Following on from the launch of OpenAI’s ChatGPT, Google’s Bard (and Duet), Microsoft’s Copilot, and X’s Grok, now Amazon has announced that it will soon be launching its own ‘Q’ generative AI chatbot (for business).
Amazon has become the latest of the tech giants to announce the introduction of its own generative AI chatbot. Recently announced at the Las Vegas conference for its AWS, ‘Q’ is Amazon’s chatbot that will be available as part of its market-leading AWS cloud platform. As such, Q is being positioned from the beginning as very much a business-focused chatbot with Amazon introducing the current preview version as: “Your generative AI–powered assistant designed for work that can be tailored to your business.”
What Can It Do?
The key point from Amazon is that Q is a chatbot that can be tailored to help your business get the most from AWS. Rather like Copilot is embedded in (and works across) Microsoft’s popular 365 apps, Amazon is pitching Q as working across many of its services, providing better navigation and leveraging for AWS customers with many (often overlapping) service options. For example, Amazon says Q will be available wherever you work with AWS (and is an “expert” on patterns in AWS), in Amazon QuickSight (its business intelligence (BI) service built for the cloud), in Amazon Connect (as a customer service chatbot helper), and will also be available in AWS Supply Chain (to help with inventory management).
Just like other AI chatbots, it’s powered by AI models which in this case includes Amazon’s Titan large language model. Also, like other AI chatbots, Q uses a web-based interface to answer questions (streamlining searches), can provide summaries, generate content and more. However, since it’s part of AWS, Amazon’s keen to show that it adds value by doing so within the context of the business it’s tailored to and becomes an ‘expert’ on your business. For example, Amazon says: “Amazon Q can be tailored to your business by connecting it to company data, information, and systems, made simple with more than 40 built-in connectors. Business users—like marketers, project and program managers, and sales representatives, among others—can have tailored conversations, solve problems, generate content, take actions, and more.” The 40 connectors it’s referring to include popular enterprise apps (and storage depositories) like S3, Salesforce, Google Drive, Microsoft 365, ServiceNow, Gmail, Slack, Atlassian, and Zendesk. The power, value, and convenience that Q may provide to businesses may also, therefore, help with AWS customer retention and barriers to exit.
Just some of the many benefits that Amazon describes Q as having include:
– Delivering fast, accurate, and relevant (and secure) answers to your business questions.
– Quickly connecting to your business data, information, and systems, thereby enabling employees to have tailored conversations, solve problems, generate content, and take actions relevant to your business.
– Generating answers and insights according to the material and knowledge that you provide (backed up with references and source citations).
– Respecting access control based on user permissions.
– Enabling admins to easily apply guardrails to customise and control responses.
– Providing administrative controls, e.g. it can block entire topics and filter both questions so that it responds in a way that is consistent with a company’s guidelines.
– Extracting key insights on your business and generating reports and summaries.
– Easy deployment and security, i.e. it supports access control for your data and can be integrated with your external SAML 2.0–supported identity provider (Okta, Azure AD, and Ping Identity) to manage user authentication and authorisation.
When, How, And How Much?
Q’s in preview at the moment with Amazon giving no exact date for its full launch. Although many of the Q capabilities are available without charge during the preview period, Amazon says It will be available in two pricing plans: Business and Builder. Amazon Q Business (its basic version) will be priced at $20/mo, per-user, and Builder at $25/mo, per-user. The difference appears to be that Builder provides the real AWS expertise plus other features including debugging, testing, and optimising your code, troubleshooting applications and more. Pricewise, Q is cheaper per month/per user than Microsoft’s Copilot and Google’s Duet (both $30).
Not All Good
Despite Amazon’s leading position in the cloud computing world with AWS, and its technological advances in robotics (robots for its warehouses), its forays in space travel (with Amazon Blue) and into delivery-drone technology, it appears that it may be temporarily lagging in AI-related matters. For example, in addition to being later to market with this AI chatbot ‘Q’, in October, a Stanford University index ranked Amazon’s Tital AI model (which is used in Q) as bottom for transparency in a ranking of the top foundational AI models with only 12 per cent (compared to the top ranking Llama 2 from Meta at 54 per cent). As Stanford puts it: “Less transparency makes it harder for other businesses to know if they can safely build applications that rely on commercial foundation models; for academics to rely on commercial foundation models for research; for policymakers to design meaningful policies to rein in this powerful technology; and for consumers to understand model limitations or seek redress for harms caused.”
Also, perhaps unsurprisingly due to Q only just being in preview, some other reports about it haven’t been that great. For example, feedback about Q (leaked from Amazon’s internal channels and ticketing systems) highlight issues like severe hallucinations and leaking confidential data. Hallucinations are certainly not unique to Q as reports about and admissions by OpenAI about ChatGPT’s hallucinations have been widely reported.
Amazon also looks like it will be makingeven greater efforts to catch up in the AI development world. For example, in September it said Alexa will be getting ChatGPT-like voice capabilities, and it’s been reported that Amazon’s in the process of building a language model called Olympus that could be bigger and better than OpenAI’s GPT-4!
What Does This Mean For Your Business?
Although a little later to the party with AI chatbot, Amazon’s dominance in the cloud market with AWS means it has a huge number of business customers to sell its business-focused Q to. This will not only provide another revenue stream to boost its vast coffers but will also enhance, add value to, and allow customers to get greater leverage from the different branches of its different cloud-related services. What with Microsoft, Google, X, Meta, and others all having their own chatbot assistants, it’s almost expected that any other big players in the tech world like Amazon would bring out their own soon.
Despite some (embarrassing internal) reviews of issues in its current preview stage and a low transparency ranking in a recent Stanford report, Amazon clearly has ambitions to make fast progress in catching up in the AI market. With its market power, wealth, and expertise in diversification and its advances in technologies like space travel and robotics and the synergies it brings (e.g. satellite broadband), you’d likely not wish to bet against Amazon making quick progress to the top in AI too.
Q therefore is less of a standalone chatbot like ChatGPT (OpenAI and former workers have helped develop AI for others) and more of Copilot and Duet arrangement in that it’s being introduced to enhance and add value to existing Amazon cloud services, but in a very focused way (more so for Builder) in that it’s “trained on over 17 years’ worth of AWS knowledge and experience”.
Despite Q still being in preview, Amazon’s ambitions to make a quantum leap ahead are already clear if the reports about its super powerful, GPT-4 rivalling (still under development) Olympus model are accurate. It remains to be seen, therefore, how well Q performs once it’s really out there and its introduction marks another major move by a serious contender in the rapidly evolving and growing generative AI market.
Following the Metro recently highlighting the issue of undisclosed cameras being used by a small number of Airbnb hosts, we take a look at what the rules say, reports in the news of this happening, and what you can do to protect yourself.
Do Airbnb Hosts Have The Right To Film Guests?
You may be surprised to know that the answer to this question is yes, hosts do have the right to install surveillance devices in certain areas of their properties (which may result in guests being filmed) but this is heavily regulated and restricted for privacy reasons.
When/Where/Why/How Is It OK For Hosts To Film Guests?
The primary legitimate reason for hosts to install surveillance devices is for security purposes. They are not allowed to use them for any invasive or unethical purposes. Airbnb’s community standards, for example, emphasise respect for the privacy of guests and any violation of these standards can lead to the removal of the host from the platform.
Airbnb’s company rules say that monitoring devices (e.g. cameras), may be used, but only if they are in common spaces (such as living rooms, hallways, and kitchens) and then only if Airbnb hosts disclose them in their listings. In short, if a host has any kind of surveillance device, they must clearly mention it in their house rules or property listing so that guests are made aware of these devices before they book the property.
What About Local Laws?
It is also the case that although disclosed cameras in common spaces on a property may be OK by the company’s rules, Airbnb hosts must also adhere to local laws and regulations regarding surveillance. This can vary widely from place to place and, in some regions, recording audio without consent is illegal, whereas video might be permissible if disclosed.
Even though Airbnb rules are relatively clear, there appears to be anecdotal and news evidence that some Airbnb guests have discovered undisclosed surveillance devices in areas of Airbnb properties where they should not be installed. Examples that have made the news include:
– Back in 2019, it was reported that a couple staying for one night at an Airbnb property in Garden Grove, California discovered a camera hidden in a smoke detector directly above the bed.
– In July 2023, a Texas couple were widely reported to have filed a lawsuit against an Airbnb owner, claiming he had put up ‘hidden cameras’ in the Maryland property they had rented for 2 nights in August 2022. According to the Court documents of Kayelee Gates and Christian Capraro, the couple became suspicious after Capraro discovered multiple hidden cameras disguised as smoke detectors in the bedroom and bathroom.
– Last month, a man (calling himself Ian Timbrell) alleged in a post on X that he had found a camera tucked between two sofa cushions at his Aberystwyth Airbnb.
Wouldn’t It Be Better To Disallow Any Cameras Inside An Airbnb Rental Property?
Banning all cameras at Airbnb rental properties might initially seem like a straightforward solution to privacy concerns, yet there are important factors to consider around this. Some hosts may legitimately need to use common areas such as entrances, for security purposes (perhaps the property is in an area where crime has been a problem) and they need to deter theft and vandalism and provide evidence if a crime occurs. On the other hand, a complete ban on cameras would address the privacy concerns of guests, ensuring they feel comfortable and secure during their stay.
Airbnb’s current policy attempts to balance security and privacy by allowing cameras in certain areas while requiring full disclosure and banning them in private spaces like bedrooms and bathrooms. However, enforcing a complete ban on cameras would be very challenging, as hidden cameras are, by nature, difficult to detect and even if there was a ban, some owners may simply not comply. The Airbnb model is built on trust between hosts and guests, and clear communication and transparency about security measures, including camera usage, are crucial for maintaining this trust. While a total ban on cameras might seem like a simple solution to privacy concerns, it overlooks the legitimate security needs of hosts. A balanced approach with clear guidelines and strict enforcement might be more effective in protecting both guest privacy and host security.
How To Check
If you’re worried about possibly being filmed/recorded by hidden and undisclosed surveillance devices in a rented Airbnb property, here are some ways you can search the property and potentially reveal such devices:
– Inspect any gadgets. Check smoke detectors or alarm clocks as they are known as places to hide cameras. Examine any other tech that seems out of place. You may also want to check the shower head.
– Search for Lenses. For example, making sure the room is dark, use a torch (such as your phone’s torch) to spot reflective camera lenses in objects like decor or appliances.
– Use phone apps like Glint Finder for Android or Hidden Camera Detector for iOS to find hidden cameras.
– Check storage areas, e.g. examine drawers, vents, and any openings in walls/ceilings.
– Check mirrors. Many people worry about the two-way mirrors with cameras behind them. Ways to check include lifting any mirrors to see the wall behind, turning off the room light and then shining a torch into the mirror to see if an area behind is visible.
– Check for infrared lights (which can be used in movement-sensitive cameras). Again, this may be spotted by by using your phone’s camera in the dark, and then looking out for any small, purple, or pink lights that may be flashing or steady.
– Scan the property’s Wi-Fi network and smart home devices for unknown devices.
– Unplug the Airbnb property’s router. Stopping the Wi-Fi at source should disable surveillance devices and may reveal whether the owner is monitoring the property, e.g. it may prompt the host to ask about the router being unplugged.
– If you’re particularly concerned, buy and bring an RF signal detector with you. Widely available online, this is a device that can find any devices emitting Bluetooth or Wi-Fi signals, e.g. wireless surveillance cameras, tracking devices and power supplies.
What Does This Mean For Your Business?
The issue of undisclosed cameras in Airbnb properties raises important considerations for Airbnb as a company, its hosts, and travellers. For Airbnb, the challenge lies in upholding and enforcing privacy standards to maintain user trust. This could involve enhancing their policies, perhaps even investing in technology or an inspection process for better detection of undisclosed devices, and/or providing more reassuring information about the issue, thereby safeguarding guest security, ensuring host accountability, and helping to protect their brand reputation.
It should be said that most Airbnb hosts abide by the company’s rules but are caught in a delicate balancing act between providing security and respecting the privacy of their guests. Any misuse of surveillance devices can, of course, have serious legal consequences and potentially harm a host’s reputation and standing on the platform. However, even just a few stories in the news about the actions on one or two hosts can have a much wider negative effect on consumer trust in Airbnb and can be damaging for all hosts. It could even simply deter people from using the platform altogether.
For some travellers, this situation may make them feel they must proactively take the responsibility for their own privacy (which may not reflect so well on Airbnb). They may feel as though they need to be informed about their rights, familiarise themselves with detection methods and remain vigilant during their stays.
This whole scenario emphasises the need for a continuous update of policies and practices by Airbnb to keep pace with technological advancements and the varying legal frameworks in different regions. It also highlights the importance of clear communication and transparency between the company, its hosts, and guests to maintain a trustworthy and secure environment.
An issue with Google’s cloud that locked some Google Drive for desktop users out of some of their files from the last six months has led to some angry comments being left on Google’s community support site.
Reported Over A Week Ago
The issue was first reported by a user (known as ‘Yeonjoong) back on 22 November. The user describes the issue (on the Google Drive Help page) as:
“The Drive literally went back to condition in May 2023. Data from May until today disappeared, and the folder structure went back to status in May. Google Drive activity doesn’t show any changes (only show activity that was in May). No files were deleted manually, so no files in Trash. I never sync or shared my files and drive to anyone, I used the drive locally.“
At the time of writing this article, Google says it is still investigating the issue. Also, the user who first reported it claims that none of the fixes suggested by Google so far have worked.
What Are Other Users Saying?
Posts from other affected users have highlighted issues such as:
– Losing access to important files from recent months.
– Questioning the dependability of the Drive Service.
– Asking for a full explanation of what had happened and to be informed about when their data will appear back.
– Fears that important data may have been altered or permanently lost, e.g. by clicking on the disconnect button (which users have now been informed not to do, but some claim were told to do originally).
– Reports of stress and worry, with some users threatening legal action.
What Does Google Advise?
At present, Google’s advice to affected Drive for desktop users in relation to what Google is calling the “Drive for desktop (v126.96.36.199 – 188.8.131.52) Sync Issue” is:
– Not to click “Disconnect account” within Drive for desktop.
– Not to delete or move the app data folder: Windows: %USERPROFILE%\AppData\Local\Google\DriveFS or macOS: ~/Library/Application Support/Google/DriveFS.
– (Optional) If users have room on your hard drive, to make a copy of the app data folder.
Google Cloud Vulnerabilities
This latest story comes hot on the heels of Bitdefender researchers reporting recently that they’d discovered vulnerabilities in Google Workspace and Google Cloud Platform which, after first compromising the local machine, could allow threat actors to extend their activities to a “chain reaction” network-wide breach, potentially leading to ransomware attacks or data exfiltration.
What Does This Mean For Your Business?
The main concern for businesses directly impacted by this issue revolves around data integrity and reliability. Losing access to recent files can disrupt ongoing projects, delay deadlines, potentially lead to financial losses, and make users very angry and frustrated (as the comments on Google’s help page show). This incident highlights the importance of having a robust backup strategy that doesn’t rely solely on cloud services. Those businesses who have been directly affected or those who may have been spooked by this story may now want to reassess their data management policies, considering additional local or multi-cloud backups for critical data.
For the wider base of Google Cloud users, this incident could be seen as a kind of cautionary tale that underscores the need for vigilance in cloud data management, and the importance of understanding the potential risks associated with cloud storage solutions. Really (time and resources permitting) users should try to stay informed about best practices for data safety and be proactive in implementing them. This could include regular audits of data access, backup strategies, and staying updated on service updates and potential vulnerabilities. That said, it seems fair for most businesses who are paying Google for aspects of its cloud service to at least expect to be able to access their files when they need them and if there is there is a problem, expect Google to sort it out quite quickly (not a week or so later). Also, many users may have been even more frustrated by a possible lack of communication on Google’s side about the issue, e.g. at least an estimate of when they could expect it to be fixed and regular updates on the situation.
For Google, this lockout issue could obviously be damaging to its reputation as a reliable cloud service provider. In the competitive cloud market, reliability and trust are paramount. Google will need to not only address the current issue swiftly and transparently but also take proactive steps to prevent similar occurrences in the future. This could involve investing more in their infrastructure, enhancing their communication protocols during crises, and possibly reviewing their update and deployment strategies to ensure minimal disruption to users. The way Google handles this situation (which many affected users haven’t been too impressed with so far) could have lasting effects on its market position and user trust.
Although this issue has posed some challenges to affected businesses, it could also be seen on reflection as providing valuable lessons for all stakeholders in the cloud services arena. Lessons include understanding more fully what their customers value the most and making more of a commitment to matters of reliability, transparency, and communication.
Analysis by the Department for Education’s Unit for Future Skills to try and quantify the impact of AI on the UK jobs market found the finance and insurance sector was more exposed than any other.
“The impact of AI on UK jobs and training” report published online by the government highlights the results of a study that used US methodology to look at the abilities needed to perform different job roles, and the extent to which these can be aided by a selection of 10 common AI applications.
These applications are:
- Abstract Strategy Games: The ability to play abstract games involving sometimes complex strategy and reasoning ability, such as chess, go, or checkers, at a high level.
- Real-time Video Games: The ability to play a variety of real-time video games of increasing complexity at a high level.
- Image Recognition: The determination of what objects are present in a still image.
- Visual Question Answering: The recognition of events, relationships, and context from a still image.
- Image Generation: The creation of complex images.
- Reading Comprehension: The ability to answer simple reasoning questions based on an understanding of text.
- Language Modelling: The ability to model, predict, or mimic human language.
- Translation: The translation of words or text from one language into another.
- Speech Recognition: The recognition of spoken language into text.
- Instrumental Track Recognition: The recognition of instrumental musical tracks.
These AI applications were selected based on their relevance and the progress in technology from 2010 onwards, as recorded by the Electronic Frontier Foundation (EFF). They represent fundamental applications of AI that are likely to have implications for the workforce and cover the most likely and most common uses of AI.
The study also focuses on which occupations, sectors and areas within the UK labour market are expected to be most impacted by AI and large language models, and how this could impact workers in different UK geographic areas.
The key findings of the study show that:
– Professional occupations are more exposed to AI, especially those associated with more clerical work and across finance, law, and business management roles.
– The industries least exposed to AI and to LLMs across industries are accommodation and food services, motor trades, agriculture, forestry, and fishing, transport and storage and construction.
– The finance and insurance sector is more exposed to AI than any other sector.
– The occupations most exposed to all AI applications are management consultants and business analysts.
– The occupations most exposed to large language modelling are telephone salespersons, followed by solicitors and psychologists.
– Workers in London and the South East have the highest exposure to AI (five times as exposed as the North-East of England), reflecting the greater concentration of professional occupations in those areas.
These findings led to some press reports that AI’s incursion into our working lives would most affect ‘city highflyers.’
Qualifications and Training
The study also exposes the qualifications and training routes that most commonly lead to these highly impacted jobs, concluding that:
– Employees with more advanced qualifications are typically in jobs more exposed to AI, e.g. those with a level 6 qualification (equivalent to a degree).
– Employees with qualifications in accounting and finance through Further Education or apprenticeships, and economics and mathematics through Higher Education are typically in jobs more exposed to AI.
Other studies highlighting levels of exposure to AI (AI taking jobs) include:
– A Pew Research Centre Study (2022) which found that 19 per cent of US workers were in jobs highly exposed to AI, where key activities might be replaced or assisted by AI.
– A Goldman Sachs Report (2023) suggesting that AI could replace the equivalent of 300 million full-time jobs globally. It indicates that about a quarter of work tasks in the US and Europe could be replaced by AI, impacting two-thirds of jobs in these regions to some degree.
A recent (October 2023) paper also highlights the dual nature of AI in advanced economies – AI’s potential as either a complement or a substitute for labour. The paper also highlights the important point that women and highly educated workers face greater occupational exposure to AI.
It’s worth noting that the Goldman Sachs Report (shown above) also highlighted this dual effect of AI, showing that AI also has the potential to create new jobs and boost productivity, potentially increasing the total annual value of goods and services produced globally by 7 per cent.
What Does This Mean For Your Business?
As highlighted in the report for this study (and as supported by the findings of other studies), 10-30 per cent of jobs are automatable with fast-evolving AI putting many of those jobs at risk. This government study confirms largely what many people may have expected – that those in more clerical work and across finance, law, and business management roles (where generative AI’s outputs are particularly effective) are most at risk from AI diminishing their value as workers. There are, of course, many other areas (some highlighted by this report) where generative AI is clearly able to replace or reproduce/copy human efforts to an acceptable degree, e.g. from customer service roles to creative work (artists). Some people may find that it’s disconcerting that jobs/professions which have taken years of study and have a specialist element and high social value (e.g. solicitors and psychologists) are shown in the report to be suddenly and significantly at risk from what are, basically, algorithms.
The report’s findings also makes what seems to be quite a logical conclusion that since there’s a greater concentration of professional occupations in London and the South East, it’s more likely to be negatively affected by AI.
The report of the study also makes the valid point about the dual nature of AI’s effects, i.e., that in addition to threatening many jobs, AI also has the potential to increase productivity and create new high value jobs in the UK economy. However, the main focus of this and other studies may appear to confirm the fears of many, that fast-advancing AI is likely to have a profound and widespread effect on the UK economy and society, and not necessarily in a good way for many peoples’ jobs, skills, and value.
As highlighted in the report, the UK education system and employers will now need to adapt to ensure that individuals in the workforce have the skills they need to make the most of the potential benefits advances in AI will bring. As individual workers, many may now want to look at the ways they can maximise their value and be in a position where they can use and orchestrate what are essentially tools more effectively than others, and in a way that adds value to themselves and their own positions, and/or in a way that creates new opportunities.
It’s been reported that iOS 17’s new NameDrop feature has prompted the Police to issue warnings to parents on social media (in the US) about its potential to be abused, possibly posing a risk to their children.
NameDrop, introduced with iOS 17.1 in November allows those with iPhones and Apple Watches running watchOS 10.1 to share their contact information, (plus photos and more with AirDrop) simply by holding their phones close together (within a few millimetres).
Police in several US states have issued warnings that this feature could, therefore, enable predators to use NameDrop to steal personal contact information from unsuspecting teens just by placing their phone next to the teen’s phone.
The Police also highlighted the fact that for protection, the feature can be turned off by going to settings, selecting AirDrop, selecting “Bringing Devices Together,” and then moving the toggle to off. There is also a ‘Contacts Ony’ option which means that only someone on the iPhone’s contact list can connect through AirDrop and NameDrop.
Researchers from cyber security firm Hunters have reported finding a Google Workspace design flaw that could allow attackers to steal emails from Gmail, data from Google Drive, and carry out other unauthorised actions within Google Workspace APIs on all of the identities in a target domain.
The design flaw (a fact reportedly disputed by Google), dubbed ‘DeleFriend,’ can be exploited by a process that involves attackers being able to leverage an existing domain-wide delegation permission to create their own fresh private key to perform API calls to Google Workspace on behalf of other identities in the domain.
It’s been reported that the Workspace domain-wide delegation feature’s potential “security risk” has been known to Google since June. Palo Alto Networks Unit 42 suggest that a way to mitigate the risk is to position service accounts with domain delegation permissions within a higher-level folder in the Google Cloud Platform (GCP) hierarchy.