A new ChatGPT tool will be able to track every time you click your mouse or tap your keyboard to help it ‘learn everything you do’.
OpenAI’s product and engineering manager Ari Weinstein said on X last week that Computer History is now available for some Apple Mac users.
By keeping an eye on your activity, the AI chatbot ‘can better understand how you work, finish tasks that you’re in the middle of and suggest skills and automations based on how you use your computer’.
How does Computer History work? Is it snooping on me? #
Computer History is an optional feature that is off by default and available only for users on the Pro plan whose desktop memory is on.
It doesn’t take screenshots or recordings of your computer screen, nor does it listen to your microphone or system audio.
Instead, ChatGPT turns all your ‘input events’, like mouse clicks, keystrokes, keyboard shortcuts and when you switch between apps, into a big timeline.
In a demo, Dominik Kundel, a member of OpenAI’s Developer Experiences team, showed the bot finding the last document he edited.
Usually, the programme would ask the user for more context before doing so.
ChatGPT double-checked if Kundel sent it to colleagues on Slack – he hadn’t, so it winged over a shareable link.
Kundel asked the model to summarise his morning, telling him that he spent it drafting and editing the video.
‘It knows what actions I took so I can converse naturally without having to give all the context,’ Kundel said.
This doesn’t mean that ChatGPT will be constantly looking over your shoulder.
OpenAI says in a feature document that you can exclude apps and websites and it does not track activity done on private web browsers.
You can also inspect and delete what the bot has memorised at any time.
The company adds that your timeline is stored on local memory files, which is data stored just on your computer.
These files are thrown into the shredder after 48 hours, but OpenAI stresses that they are not encrypted.
‘Other programmes running as your macOS user may be able to access them,’ it says.
‘Protect your Mac account and exclude sources you do not want included.’
OpenAI does briefly get a hold of your files so it can process them into ‘memories’, data that the chatbot can remember.
‘OpenAI does not retain those event files after processing unless required by law and does not use them for training,’ the document adds.
‘Just another valuable source of information to steal’ #
The worry, cyber security experts told Metro, isn’t that a hacker breaks into your ChatGPT and discovers you use it to write wedding speeches.
Many cyber criminals trick victims into down malware, shady software which can sneakily take control of a device and steal data.
‘If an attacker is already running malware on your system, these unencrypted summaries are just another valuable source of information to steal,’ Stefanie Schappert, cybersecurity expert at Cybernews, says, ‘potentially exposing the very health, financial, and personal information OpenAI explicitly warns users to exclude from their Computer History.
‘In the wrong hands, that information could then be used for targeted phishing and social engineering attacks, credential and identity theft, corporate espionage, or further account compromise.’
But Wade Woolwine, senior director of product security at the cybersecurity platform Rapid7, says OpenAI has given users a lot of control over what Computer History can see and do.
‘I don’t think users who opt into this feature are significantly increasing the risk of private information being disclosed,’ he says.
‘An attacker would first need to compromise the user’s computer. Even if the user has the feature disabled, that same type of data would be accessible from any number of other logs on the system.’
One concern that OpenAI acknowledges in its Computer History document is prompt injection risk.
AI bots want to make users happy, so the tech sometimes carries out a request, called a prompt, even if it might be a malicious one.
So OpenAI says that there is an ‘increased risk’ that ChatGPT may follow malicious prompts hidden inside dodgy apps or websites.
A chatbot that doesn’t need to be endlessly reminded about the user’s life makes sense, Alex Goller, principal solution architect EMEA at the cyber breach containment firm Illumio, says.
It’s even been done before – Microsoft’s Recall, an AI feature that takes snapshots of mostly everything you see on your computer screen, was introduced last year.
But experts criticised the lack of an option to disable Recall at the time and worried just how much it would snoop on people.
‘OpenAI has learned a lot from that situation – there are no constant screenshots here, it’s opt-in and private browsing is excluded,’ Goller says.
‘As AI adoption matures, we’re all focusing on what in our day-to-day work is worth automating and no model can answer that without knowing how a user actually works.
‘This is why OpenAI has created this new feature and this kind of activity capture is where the whole AI industry is heading, which is exactly why the security bar has to be set now.’
Get in touch with our news team by emailing us at webnews@metro.co.uk.
**For more stories like this, **[ check our news page](https://metro.co.uk/news/).
MORE: [Teenager charged with killing mum and brother ‘used ChatGPT to play out fantasy’](https://metro.co.uk/2026/08/17/boy-17-charged-killing-mum-brother-used-chatgpt-play-fantasy-29396297/?ico=more_text_links)
MORE: [Games Inbox: When will The Elder Scrolls 6: Sentinel be released?](https://metro.co.uk/2026/08/13/games-inbox-will-elder-scrolls-6-sentinel-released-29365052/?ico=more_text_links)
MORE: [Saber admit they use AI while denying they swapped lead writer for ChatGPT](https://metro.co.uk/2026/08/12/saber-admit-use-ai-denying-swapped-lead-writer-chatgpt-29359907/?ico=more_text_links)