Why is Google Bard introduced in 180 countries but not in Netherlands, France, Germany, Spain or any other EU country? In the next few tweets I will try to give a possible explanation.
The potential compliance issues with upcoming "AI regulations" in Europe could be a concern for large language models like Google Bard. These proposed regulations are still in a nebulous stage and have not been enacted into law. The European Union's perspective is that AI should⦠https://twitter.com/i/web/status/1656947949794996224
Reason 1 why Bard is not yet in EU. Understanding how AI makes decisions is crucial. The EU's proposed regulations emphasize this, but it's a tough task for Large Language Models (LLMs). Their complex decision-making process isn't easily interpretable, making full transparency a⦠https://twitter.com/i/web/status/1656948300522614785
2. Risk to Fundamental Rights
If an LLM is used in a way that could pose a risk to people's fundamental rights, such as privacy or non-discrimination, it might not comply with the regulation. For example, if an LLM is trained on data that includes biased or discriminatory web⦠https://twitter.com/i/web/status/1656949622726619136
3. Data Governance
Quality data is key to reliable AI. EU's proposed AI regulations stress this. But if an LLM is trained on a dataset with errors or biases, it could lead to unreliable outputs. Good data governance is essential for trustworthy AI. #AI #DataGovernance https://twitter.com/i/web/status/1656949890092597253
4. Human Oversight
Human oversight is vital in AI systems. The EU's proposed AI regulations require this, especially for high-risk AI systems. Depending on how an LLM is used, it might need more robust human oversight mechanisms. Let's keep humans in the AI loop. #AI⦠https://twitter.com/i/web/status/1656950156812533760
5. Current law
Google Bard and similar AI systems could already now potentially run afoul of EU and Canadian regulations on data protection, intellectual property, and consumer rights. Ensuring AI compliance with these rules is essential for safeguarding our digital rights #ai https://twitter.com/i/web/status/1656950732765945859
Check https://support.google.com/bard/answer/13575153?hl=en to find out which countries have access. EU is still blocked as we speak. (END)
Google Bard and similar LLM's can already run afoul of EU, Canadian or Brazilian regulations on existing data protection. Here is a completely fake bio of me and my friend in search Paul Meyers (BBC) when I asked Bard: how did we meet? Nothing of the personal story is true. https://t.co/dQqXpUlzd9
Under the European GDPR, individuals have the right to have their personal data rectified if it is inaccurate or incomplete. Article 16 of the GDPR grants individuals the right to request the rectification of their personal data from the data controller. If an individual believes⦠https://twitter.com/i/web/status/1657476032469893123 https://t.co/mCXIB11zj0
Although I misspelled the name (correct one is @PaulMyersBBC ), Bing got it right. Bard is spitting out more nonsense, see screenshot.
Both Paul and I are unable to rectify the bot's behavior.
As for a Bard gatekeeper, it does not currently exist. Google is aware of this fact and⦠https://twitter.com/i/web/status/1657479105221197828 https://t.co/iwAMAGMwCF
Rewrote https://www.digitaldigging.org/p/why-is-google-bard-blocked-in-eu based on this thread, completely. Did that in this bus :) https://t.co/AQDBpdVnCb
Also, have a look at https://twitter.com/henkvaness/status/1657987337064775680?s=20 - the full proof that I am a professional banjo player from Saint Kitts.