Factbox: Governments race to regulate AI tools

Melco International Development

28/11/2023 11:52

(Updates the U.S.)
       Nov 28 (Reuters) - Rapid advances in artificial
intelligence (AI) such as Microsoft-backed OpenAI's ChatGPT are
complicating governments' efforts to agree laws governing the
use of the technology.
    Here are the latest steps national and international
governing bodies are taking to regulate AI tools:

    AUSTRALIA
    * Planning regulations
    Australia will make search engines draft new codes to
prevent the sharing of child sexual abuse material created by AI
and the production of deepfake versions of the same material.

    BRITAIN
    * Planning regulations
    Leading AI developers agreed on Nov. 2, at the first global
AI Safety Summit in Britain, to work with governments to test
new frontier models before they are released to help manage the
risks of the developing technology. 
    More than 25 countries present at the summit, including the
U.S. and China, as well as the EU, on Nov. 1 signed a "Bletchley
Declaration" to work together and establish a common approach on
oversight.
    Britain said at the summit it would triple to 300 million
pounds ($364 million) its funding for the "AI Research
Resource", comprising two supercomputers which will support
research into making advanced AI models safe, a week after Prime
Minister Rishi Sunak had said Britain would set up the world's
first AI safety institute.
    Britain's data watchdog said in October it had issued Snap
Inc's  SNAP.N  Snapchat with a preliminary enforcement notice
over a possible failure to properly assess the privacy risks of
its generative AI chatbot to users, particularly children.

    CHINA
    * Implemented temporary regulations
    Wu Zhaohui, China's vice minister of science and technology,
told the opening session of the AI Safety Summit in Britain on
Nov. 1 that Beijing was ready to increase collaboration on AI
safety to help build an international "governance framework".
    China published proposed security requirements for firms
offering services powered by generative AI in October, including
a blacklist of sources that cannot be used to train AI models.
    The country issued a set of temporary measures in August,
requiring service providers to submit security assessments and
receive clearance before releasing mass-market AI products.
 
    EUROPEAN UNION
    * Planning regulations
    France, Germany and Italy have reached an agreement on how
AI should be regulated, according to a joint paper seen by
Reuters on Nov. 18. The paper explains that developers of
foundation models would have to define model cards, which are
used to provide information about a machine learning model.
    European lawmakers agreed on Oct. 24 on a critical part of
new AI rules outlining the types of systems that will be
designated "high risk", inching closer to a broader agreement on
the landmark AI Act which is expected in December, according to
five people familiar with the matter.

    FRANCE
        * Investigating possible breaches
    France's privacy watchdog said in April it was investigating
complaints about ChatGPT.

    G7
        * Seeking input on regulations
    The Group of Seven countries agreed on Oct. 30 to an
11-point code of conduct for firms developing advanced AI
systems, which "aims to promote safe, secure, and trustworthy AI
worldwide". 

    ITALY
        * Investigating possible breaches
    Italy's data protection authority plans to review AI
platforms and hire experts in the field, a top official said in
May. ChatGPT was temporarily banned in the country in March, but
it was made available again in April.

    JAPAN
        * Investigating possible breaches
    Japan expects to introduce by the end of 2023 regulations
that are likely closer to the U.S. attitude than the stringent
ones planned in the EU, an official close to deliberations said
in July.
    The country's privacy watchdog has warned OpenAI not to
collect sensitive data without people's permission.

    POLAND
        * Investigating possible breaches
    Poland's Personal Data Protection Office said in September
it was investigating OpenAI over a complaint that ChatGPT breaks
EU data protection laws.

    SPAIN
    * Investigating possible breaches
    Spain's data protection agency in April launched a
preliminary investigation into potential data breaches by
ChatGPT.

    UNITED NATIONS
    * Planning regulations
    The U.N. Secretary-General António Guterres on Oct. 26
announced the creation of a 39-member advisory body, composed of
tech company executives, government officials and academics, to
address issues in the international governance of AI.

    UNITED STATES
    * Seeking input on regulations
    The U.S., Britain and more than a dozen other countries on
Nov. 27 unveiled a 20-page non-binding agreement carrying
general recommendations on AI such as monitoring systems for
abuse, protecting data from tampering and vetting software
suppliers.
    The U.S. will launch an AI safety institute to evaluate
known and emerging risks of so-called "frontier" AI models,
Secretary of Commerce Gina Raimondo said on Nov. 1 during the AI
Safety Summit in Britain.
    President Joe Biden issued an executive order on Oct. 30 to
require developers of AI systems that pose risks to U.S.
national security, the economy, public health or safety to share
the results of safety tests with the government.
    The U.S. Federal Trade Commission opened in July an
investigation into OpenAI on claims that it has run afoul of
consumer protection laws.

 (Compiled by Alessandro Parodi and Amir Orusov in Gdansk;
Editing by Milla Nissi and Emelia Sithole-Matarise)
 ((alessandro.parodi@thomsonreuters.com))