Ask a simple question in Ghana today, and the answer surprises most people. Does Ghana have a law about artificial intelligence, known as AI? The short answer is no.
Ghana launched a National AI Strategy on 24 April 2026. President John Dramani Mahama launched it in Accra, and the government committed $270 million to it. That is a plan, not a law. A strategy says what a country wants to do. A law says what people and companies must do, and what happens if they do not.
So what rules apply right now when a bank uses AI to decide on a loan, when a company uses AI to screen job applicants, or when someone uses AI to make a fake video of a real person? The rules come from older laws. Those laws were written before AI became common, and they cover only part of the problem. This article explains each law in simple terms, what it covers, what it does not cover, and what is being worked on next.
What Is Actually Happening
AI is already in use in Ghana, in ways most people do not see. A bank may use AI to spot a strange transaction and block it. A lender may use AI to help decide if a person can repay a loan. A shop may use AI to answer customer messages. A school may use AI to mark work or to plan lessons. A person with a phone can use AI to write a message, make a picture, or copy a voice.
The National AI Strategy, which runs from 2025 to 2035, is built around eight areas: AI education, youth employment, digital infrastructure, data governance, support for new businesses, AI use in sectors such as farming, health and public services, research, and AI use in government. The government committed $270 million in total. Of that, $250 million is for a national AI computing centre, a place with powerful computers that Ghanaian researchers and businesses can use. The other $20 million is for short to medium term work on the strategy. A National AI Fund is planned, starting at GH₵5 billion over its first five years. A Responsible AI Office is planned under the Ministry of Communication, Digital Technology and Innovations, to coordinate the work and watch progress.
None of that creates a new legal duty by itself. The strategy does not give a person a right to see why an AI system said no to them. It does not set a fine for a company that uses AI unfairly. It does not make it a crime to make a deepfake video. Those things need a law, and that law does not exist yet.
That gap matters because AI is moving faster than the law. The same month the strategy was launched, the Cyber Security Authority was dealing with AI deepfake videos that pretended to be the President. We explain that case in our article about the deepfake president videos. A fake video can be made in minutes. A new law takes years.
The Laws That Apply Today
Four sets of rules do most of the work today. None of them is an AI law, but each one reaches part of what AI does.
1. The Data Protection Act, 2012 (Act 843)
This is the main law about personal data in Ghana. Personal data is information that can identify a person, such as a name, a phone number, a photo, or a voice recording. The law is run by the Data Protection Commission, known as the DPC.
The Act says a person or company that collects personal data must have a lawful reason, must use it only for the purpose they stated, must keep it accurate and safe, and must register with the DPC before they start. Those rules apply to AI systems too. If an AI system uses your photo, your voice, or your bank details, the company behind it must follow the same rules as any other company that holds your data.
The limit is clear. The Act was passed in 2012, long before the AI tools people use today. It does not use the words artificial intelligence, algorithm, or automated decision. It does not give you a clear right to ask why an AI system made a decision about you, or to ask a human to review that decision. As of October 2026, the DPC had not issued AI specific rules. That leaves a grey area, a space where the law is not clear, for anyone whose loan, job application, or account was decided by a machine.
2. The Cybersecurity Act, 2020 (Act 1038)
This Act set up the Cyber Security Authority, known as the CSA. It covers cybercrime, which is crime committed with a computer or a phone, and it covers the protection of important computer systems, such as those used by banks and by government.
The Act is the reason the CSA can investigate online fraud, including fraud that uses AI. It is also the law behind the licensing of cybersecurity companies. In August 2026, the CSA fined EY Ghana GH₵360,000 for providing cybersecurity services without a licence. We covered that case in our article about the CSA fine on EY Ghana.
The Act does not mention AI by name either. It was not written to cover a fake video, a cloned voice, or an AI system that makes a biased decision. Reports in 2026 said the CSA was seeking changes to the Act to make it a crime to produce a deepfake with the aim of deceiving people. As of November 2026, no such change had been passed into law.
3. The Electronic Communications Act, 2008 (Act 775)
This Act covers phones, the internet, and broadcasting in Ghana. It includes rules on consumer protection, which is the protection of people who buy or use a service, and on the duties of the companies that provide phone and internet services.
It matters for AI in two ways. First, the National Communications Authority, known as the NCA, regulates the networks that AI services run on. Second, some of the consumer protection rules in the Act can apply when a company misleads a customer, including with an AI generated advert. But the Act was written for phones and radio, not for AI, and it does not set rules for how AI systems should be built or tested.
4. The Bank of Ghana rules for banks and fintechs
The clearest AI rules in Ghana today are in banking, and they come from the Bank of Ghana, not from Parliament.
In March 2026, the Bank of Ghana launched a revised Cyber and Information Security Directive, known as CISD 2026. It replaces the 2018 version. It applies to banks, microfinance firms, savings and loans companies, fintechs, and payment firms that the Bank of Ghana regulates. A fintech is a company that uses technology to offer money services, often on a phone.
CISD 2026 has six main areas. The first is AI and machine learning governance. Machine learning is a type of AI that learns from data. Banks must show that their AI systems are open about how they work, treat customers fairly, and are kept safe. That covers AI used for fraud detection, credit scoring, and customer service. Credit scoring is the process of judging whether a person is likely to repay a loan.

We explain CISD 2026 in full in our article Beyond the Banks: What Ghana’s Cyber and Information Security Directive 2026 Means for the Wider Digital Economy.
The Bank of Ghana is also working on a separate directive about AI use in the financial sector. In October 2026, the Governor said the central bank was developing AI rules for banks and fintechs, naming credit scoring, fraud detection, and customer service as the main uses. No draft text had been published as of October 2026, so businesses cannot yet read the exact rules. Until it appears, CISD 2026 is the main written rule for AI in Ghanaian banking.
Why This Is Different
Other countries have already passed AI laws. The European Union passed its AI Act in 2024. It sorts AI systems by risk, with strict rules for the highest risk uses, such as deciding on jobs or loans, and bans on some uses, such as certain types of face recognition in public. Ghana has not taken that path yet. Its approach so far is a strategy plus sector rules, starting with banking.
That choice has a clear benefit. A strategy can move fast, and sector rules can be written by the regulator that knows the sector best. The Bank of Ghana knows banking. The CSA knows cybersecurity. A single AI law, written for every sector at once, would take longer and could easily get parts wrong.
It also has a clear cost. The protection you get depends on which sector you are in. A bank customer has CISD 2026 behind them. A job applicant screened by AI, a student marked by AI, or a patient whose details were fed into an AI tool has much less. The same AI system can be governed in one setting and ungoverned in another, with no single office to complain to.
The second cost is about proof. When a human makes a decision, you can ask them why. When an AI system makes a decision, the reason may sit inside a model that the company bought from abroad, or that it cannot explain in simple terms. Ghanaian law today does not give you a clear right to that explanation. Our article about who is liable when AI goes wrong in Ghana explains this gap in detail, and why it matters most for banks, lenders, insurers, and employers.
The third cost is about deepfakes. A fake video of a real person is not clearly a crime under a law written for AI, because no such law exists. It can be charged as fraud, if money was taken by false pretences, or under the Cybersecurity Act, if it was part of a cybercrime. But the act of making the fake, on its own, sits in that grey area. That is why the CSA has sought a change to the law.
Who Is Actually at Risk
Ordinary people are at risk in ways that are easy to miss, because the AI is often invisible.
Borrowers are at risk first. If a loan app uses AI to decide who gets a loan, and the system has learned from old data that favours one group over another, some people will be refused for reasons they cannot see and cannot challenge. There is no clear right today to ask for the reason, or to ask a human to look again.
Job seekers are at risk too. Some employers use AI to screen CVs or to score video interviews. A system trained on past hiring decisions can repeat past bias, and the applicant may never know that a machine, not a person, rejected them.
Small businesses are at risk in a different way. A shop owner who pastes customer names, phone numbers, or payment details into a free AI chat tool may break the Data Protection Act without knowing it. The Act applies even when the tool is free, and even when the owner did not mean to share the data. Our guide to closing Ghana’s AI adoption gap in banking shows how far most Ghanaian firms still are from having clear AI rules of their own.
Anyone with a face or a voice online is at risk from deepfakes. A few seconds of voice from a video is enough to clone it. The May 2026 arrests for deepfake videos of the President show that the risk is real, and that it reaches the highest office in the country, not only ordinary people.

What Is Being Worked On
Three pieces of work could close part of the gap, but none of them is law yet.
First, a new Data Protection Bill, dated 2025, has been drafted and put out for consultation. It would replace Act 843. In February 2026, the Minister for Communication said the new bill was being developed to cover AI, automated decisions, and the movement of data across borders. A companion bill on emerging technologies, covering AI, advanced data analysis, digital assets, and new digital platforms, was also announced at the same time. As of September 2026, Parliament’s bills register listed no Data Protection Bill, so the bill had not yet been formally tabled. Until it is passed, Act 843 remains the law.
Second, the Bank of Ghana AI directive, described above, is still being developed. When it is published, it will be the first set of AI rules written for one sector in full, and other regulators may follow its pattern.
Third, the National AI Strategy itself promises further regulation over its ten year life, to 2035. A strategy can promise, but only Parliament can pass a law. The timing of any AI bill will depend on when the government puts one before Parliament, and that date has not been announced.
What Ghanaian Law Says, in Short
Ghana has no law with artificial intelligence in its name. The rules that apply today are the Data Protection Act 2012 (Act 843), the Cybersecurity Act 2020 (Act 1038), the Electronic Communications Act 2008 (Act 775), and, for banks and fintechs, the Bank of Ghana’s CISD 2026. A new Data Protection Bill and a Bank of Ghana AI directive are both being worked on. Neither has been passed or published yet.
That is general information, not legal advice. If you face a real legal decision, such as a refused loan, a lost job, or a dispute with a company over your data, speak to a qualified Ghanaian legal practitioner.
What to Actually Do: A Checklist

Use this checklist whether you are a person using AI, or a business thinking of using it.
- Ask what the AI does with your data. Before you use an AI tool for work, ask what data it keeps, where it is stored, and whether it is used to train the system. If the tool cannot answer, do not paste personal details into it.
- Never paste customer details into a free AI chat tool. Names, phone numbers, payment details, and health information belong in your own systems, not in a public AI chat. Under Act 843, you are responsible for that data.
- Ask for a human review when a machine says no. If a loan, a job, or an account decision was made by a machine, ask in writing for the reason and for a human to review it. The law does not yet guarantee this, but asking creates a record, and a fair company will answer.
- Check before you trust a video or a voice note. If a video shows a known person asking for money or selling a product, check it on a separate channel you already trust, such as the person’s official page or a known phone number. Do not trust a familiar face or voice alone.
- Keep proof. Save the message, the screenshot, the name of the app, and the date. If you need to report it, that proof is what the CSA and the police will ask for. Report scams to the CSA by calling or texting 292, sending a WhatsApp message to 0501603111, or emailing report@csa.gov.gh.
- If you run a business, write down your AI rules now. List which tools your staff may use, what data they may never paste into them, and who checks the output before it goes to a customer. A one page rule, written now, is better than a policy written after a problem.
Conclusion
Ghana has a strategy, a computing centre in the works, and $270 million committed. What it does not yet have is a law about AI. The rules that apply today were written for data protection, for cybersecurity, and for phones, and they were written before AI became common. They cover part of the problem, and they leave clear gaps: no clear right to an explanation when a machine says no, no clear crime for making a deepfake, and no single office to complain to.
The strategy runs to 2035. The law will need to catch up long before then, because AI is already deciding, screening, and faking in Ghana today. Until it does, the best protection is the simple one: ask what the AI does with your data, ask for a human to review a machine decision, and check a familiar face before you trust it with your money.
Sources Referenced
- National AI Strategy 2025 to 2035, launched 24 April 2026 by President John Dramani Mahama, $270 million committed, $250 million for a national AI computing centre and $20 million for short to medium term work, eight pillars, Responsible AI Office planned. Reported by Citi Newsroom and others, April 2026. https://chale.tech/ghana-ai-compute-centre/
- Ghana has no dedicated AI law as of April 2026, and no AI bill tabled in Parliament. Existing rules are the Data Protection Act 2012 (Act 843), the Cybersecurity Act 2020 (Act 1038), and the Electronic Communications Act 2008 (Act 775). https://www.jbklutse.com/ghana-ai-law-2026/
- Ghana has no AI law as of 29 October 2026. A new Data Protection Bill and a Bank of Ghana AI directive are both still being prepared. Bank of Ghana Governor said in October 2026 the central bank is developing a directive on AI use in the financial sector, naming credit scoring, fraud detection, and customer service. https://www.zigmaneural.com/ai-in/ghana
- Data Protection Bill 2025 drafted and consulted, not listed in Parliament’s bills register as of 10 September 2026. Minister said in February 2026 a new bill was being developed to cover AI, automated decision making, and cross border data flows, and an Emerging Technologies Bill was also announced. https://www.recordinglaw.com/world-laws/world-data-privacy-laws/ghana-data-privacy-laws/
- Bank of Ghana, Cyber and Information Security Directive 2026, March 2026, six pillars including AI and machine learning governance, replacing the 2018 directive. https://techlabari.com/ghanas-new-cybersecurity-rules-will-force-banks-to-rethink-information-security/
- Cyber Security Authority: https://www.csa.gov.gh/
- Bank of Ghana: https://www.bog.gov.gh/
- European Union AI Act, Regulation (EU) 2024/1689, passed 2024, risk based rules for AI systems. https://eur-lex.europa.eu/eli/reg/2024/1689/oj

