Anthropic opens model access to METR for independent AI safety review
Amid wider scrutiny of how leading AI labs manage outside oversight, Anthropic says it is giving nonprofit evaluator METR employee-level access to its systems for independent reviews. The arrangement is intended to let METR examine the company's advanced models, safety practices and risk management with access comparable to Anthropic's internal assessment teams.
Highlights
- Anthropic granted six METR employees access to company hardware, office space, and internal tools matching its own risk assessment teams’ resources.
- METR can independently assess and publicly report on Anthropic's advanced AI models and safety practices, constrained only by limited redactions for sensitive data.
- Anthropic positions this arrangement as an AI sector governance measure amid rising scrutiny of model safety and concerns about developer influence over independent reviews.
Independent review access and scope
As reported by Reuters, Anthropic says six METR employees have been provided company hardware, office space and access to internal tools and systems that are broadly comparable to those available to the company's own risk assessment teams.The company says the access allows METR to independently assess advanced models, safety practices and risk management inside the AI lab. Anthropic also says METR will be able to publish independent assessments of risks, incidents and safety practices without prior approval, subject to limited redactions tied to customer data, privileged information and certain model architecture details.
Implications for AI governance
The move follows concerns among researchers and evaluators that leading AI developers could exert too much influence over independent reviews through publication restrictions and control over access to sensitive information.By extending employee-level access to an outside nonprofit evaluator, Anthropic is positioning the arrangement as a governance measure for the AI sector, where scrutiny of model safety and internal risk controls is intensifying.
In our earlier coverage of the debate over managing AI risks, we highlighted how governments and major developers are exploring stronger safeguards and international coordination as advanced systems accelerate. The article noted proposals such as a U.S.-China incident communication channel and discussions about using the UK as a neutral venue to align on AI safety standards and prevent misuse.
Latest Company News News
-
Afghanistan
-
Albania
-
Algeria
-
Angola
-
Argentina
-
Armenia
-
Australia
-
Austria
-
Azerbaijan
-
Bahamas
-
Bahrain
-
Bangladesh
-
Belarus
-
Belgium
-
Bolivia
-
Botswana
-
Brazil
-
Brunei
-
Bulgaria
-
Cambodia
-
Cameroon
-
Canada
-
Chile
-
China
-
Colombia
-
Congo
-
Costa Rica
-
Cote d'Ivoire
-
Croatia
-
Cuba
-
Cyprus
-
Czechia
-
DR Congo
-
Denmark
-
Dominican Republic
-
Ecuador
-
Egypt
-
El Salvador
-
Estonia
-
Eswatini
-
Ethiopia
-
Finland
-
France
-
Georgia
-
Germany
-
Ghana
-
Greece
-
Haiti
-
Hong Kong
-
Hungary
-
India
-
Indonesia
-
Iran, Islamic republic
-
Iraq
-
Ireland
-
Israel
-
Italy
-
Jamaica
-
Japan
-
Jordan
-
Kazakhstan
-
Kenya
-
Korea
-
Kuwait
-
Kyrgyzstan
-
Laos
-
Latvia
-
Lebanon
-
Lesotho
-
Libya
-
Lithuania
-
Luxembourg
-
Madagascar
-
Malaysia
-
Malta
-
Mauritius
-
Mexico
-
Moldova
-
Mongolia
-
Montenegro
-
Morocco
-
Mozambique
-
Myanmar
-
Namibia
-
Nepal
-
Netherlands
-
New Zealand
-
Nigeria
-
North Macedonia
-
Norway
-
Oman
-
Pakistan
-
Palestine
-
Panama
-
Papua New Guinea
-
Paraguay
-
Peru
-
Philippines
-
Poland
-
Portugal
-
Puerto Rico
-
Qatar
-
Reunion
-
Romania
-
Rwanda
-
Saudi Arabia
-
Serbia
-
Singapore
-
Slovakia
-
Slovenia
-
Somalia
-
South Africa
-
Spain
-
Sri Lanka
-
Sweden
-
Switzerland
-
Syria
-
Taiwan
-
Tajikistan
-
Tanzania
-
Thailand
-
Trinidad and Tobago
-
Tunisia
-
Turkey
-
UAE
-
Uganda
-
Ukraine
-
United Kingdom
-
United States
-
Uruguay
-
Uzbekistan
-
Venezuela
-
Vietnam
-
Yemen
-
Zambia
-
Zimbabwe
- Forex
- Crypto
