The UK is falling behind in protecting its people from dangerous AI-generated misinformation, and the Online Safety Act is doing little to mitigate the damage, a new report from Full Fact claims.

With the next general election looming, quelling the spread of misinformation and disinformation is increasingly vital, especially as emerging technology like generative AI makes it even harder for people to tell what is real and what is fake.

The World Economic Forum (WEF) said that misinformation and disinformation will be the top short-term threats of our time, largely driven by advancements in AI.

The rise of AI has improved the ability of threat actors and fraudsters to create genuine-looking misinformation and spread it like wildfire online, with deepfakes proliferating across social media impersonating celebrities and politicians alike.

Classic misinformation, such as false statistics, are still apparent as well, and the UK government appears to be failing at filtering both.

The fact-checking charity found several instances where UK politicians did not correct false or misleading claims despite Full Fact calling on them to do so, and found that the statistics regulator had to write to the UK government at least ten times in the past year to challenge its use of statistics or other data.

The Online Safety Act, which passed just last year, leaves several gaps in combatting misinformation online, Full Fact claims.

While the Act does stipulate against harmful content, it does not have specific guidance or regulation geared to harmful misinformation outwith normal libel laws. It focuses primarily on content that is harmful in a physical and psychological sense, rather than content that is harmful to society and democracy.

Mitigating misinformation, then, is still largely left up to platforms, which can set up their own set of guidelines to combat, remove, flag, or keep up misleading content.

Recommended reading

• Google Restricts Gemini From Answering 2024 Election Queries
• Gemini Art Generator Paused Over Woke-ism Allegations
• ChatGPT to the Left, Meta’s LLaMA to the Right: Stuck in the Middle of AI Bias

Further, Full Fact’s report calls out the UK government’s lackluster and already out of date position on AI regulation ahead of the election, calling out the UK’s 2022 White Paper that does not address generative AI and its associated risks.

“In the multi sector and jurisdiction space like generative AI we need citizen-supporting initiatives on a new scale and opportunities for the public and civil society to shape related policy in the UK, including on misinformation and disinformation,” Full Fact’s report said.

It is calling on the UK to explicitly approach generative AI and its associated risks, and form its conclusive steps in consultation with experts and civil society.

In an effort to combat against AI-generated misinformation, Google, the world’s largest search engine, has taken some steps to label AI-generated content on YouTube, and has even said that its AI chatbot, Gemini, will not answer any questions related to upcoming global elections in the UK, India, and the US.