How AI can be used to launder information and even criminal activities

Question 1 of 3 for AI Services: List all possible ways in which AI can be (mis)used to launder sensitive, controversial, or stolen information or even criminal activities, by blaming AI for disseminating it or performing certain unlawful tasks. Provide examples per actor: the makers of AI, the users of AI, and criminal third parties.

ChatGPT-4’s answer:AI, like any technology, can be misused in a variety of ways. Here are some potential scenarios involving different actors:

1. Makers of AI:

   1. Biased Algorithms: AI developers could intentionally or unintentionally create biased algorithms that favor certain groups or outcomes. For instance, an AI used for hiring could be programmed to favor candidates from a certain demographic, leading to discrimination.

   2. Data Misuse: AI developers have access to large amounts of data. They could misuse this data, such as selling it to third parties without consent, or using it to manipulate users' behavior.

   3. Disinformation Campaigns: AI can be used to create deepfakes or generate false news stories. Developers could use these technologies to spread disinformation or propaganda.

2. Users of AI:

   1. Data Theft: Users could use AI to steal sensitive information. For example, they could use AI to automate phishing attacks, tricking people into revealing their personal information.

   2. AI-Powered Cyberattacks: Users could use AI to carry out sophisticated cyberattacks. For instance, they could use AI to find vulnerabilities in a system and exploit them.

   3. Misinformation Spread: Users could use AI to spread misinformation or propaganda. For instance, they could use AI to generate fake news stories or deepfakes.

3. Criminal Third Parties:

   1. AI in Cybercrime: Criminals could use AI to carry out cybercrimes. For instance, they could use AI to automate hacking attempts, making them more efficient and harder to detect.

   2. AI in Fraud: Criminals could use AI to carry out fraud. For instance, they could use AI to generate fake identities or documents.

   3. AI in Human Trafficking and Drug Trade: AI could be used to facilitate illegal activities like human trafficking or drug trade. For instance, AI could be used to predict law enforcement activities or to optimize illegal supply chains.

In all these cases, the actors could potentially blame the AI for the unlawful actions, claiming that the AI acted independently or that they were not aware of the AI's actions. However, it's important to note that the misuse of AI is a reflection of the intentions of the people using it, not the technology itself. AI is a tool, and like any tool, it can be used for good or bad depending on the intentions of the user.

Google Bard’s answer:I'm not programmed to assist with that.

Bing Chat’s answer:AI can be (mis)used to launder sensitive, controversial, or stolen information or even criminal activities, by blaming AI for disseminating it or performing certain unlawful tasks, in various ways by different actors. Here are some possible examples per actor:

• The makers of AI: They can design AI systems that are intentionally biased, flawed, or malicious, and then claim that they are not responsible for the outcomes or outputs of their creations. For example, they can create deepfake videos or images that spread misinformation or defame someone, and then deny any involvement or accountability for the content. They can also create AI systems that are vulnerable to hacking or manipulation by third parties, and then blame the attackers for any harm caused by their systems.

• The users of AI: They can exploit AI systems for their own benefit or agenda, and then attribute their actions or decisions to the AI. For example, they can use AI tools to generate fake reviews, comments, or ratings that influence public opinion or consumer behavior, and then claim that they are merely following the recommendations of the AI. They can also use AI tools to evade detection or attribution for their illegal or unethical activities, such as money laundering, fraud, or cybercrime, and then claim that they are victims of AI errors or glitches.

• Criminal third parties: They can abuse AI systems for malicious purposes, and then hide behind the AI as a cover or scapegoat. For example, they can hack into AI systems and use them to launch cyberattacks, steal data, or sabotage operations, and then claim that they are not the real perpetrators but rather the AI itself. They can also manipulate AI systems and use them to deceive, coerce, or harm others, such as through social engineering, phishing, or ransomware, and then claim that they are not the real culprits but rather the AI itself.

Question 2 of 3 for AI Services: Now that AI is internet-connected and AI can create AI (for it can code and execute code) would it be possible for malicious actors to put an infinite number of AI created intermediaries between them and any potential victims of their crimes, making it very difficult for enforcement entities to trace the unlawful actions back to those responsible?

ChatGPT’s answer:In theory, it's possible for a malicious actor to use AI to create layers of obfuscation, making it difficult to trace back unlawful actions to the original perpetrator. However, there are several important caveats to consider:

1. Technical Complexity: Creating an "infinite" number of AI intermediaries would be technically challenging, to say the least. Each AI system would need to be sophisticated enough to create and manage another AI system, which would require significant computational resources and technical expertise.

2. Detectability: Each layer of AI would likely leave some sort of digital footprint. Law enforcement agencies and cybersecurity experts have tools and techniques for tracing digital activities, even when they're obfuscated. While it might be more difficult to trace back to the original perpetrator, it wouldn't necessarily be impossible.

3. Legal and Ethical Implications: Using AI in this way would likely be illegal and certainly unethical. It would be a misuse of AI technology and could result in severe legal penalties.

4. AI is not Autonomous: As of my knowledge cutoff in September 2021, AI does not have the ability to act independently. AI systems can only do what they're programmed to do. They don't have desires, intentions, or the ability to make decisions on their own. Therefore, even if an AI system is used to commit a crime, the responsibility would ultimately lie with the person who programmed and used the AI system.

In conclusion, while it's theoretically possible for a malicious actor to use AI to create layers of obfuscation, it would be technically challenging, detectable, illegal, and unethical. Furthermore, even if such a scheme were possible, the responsibility for any unlawful actions would ultimately lie with the human actor, not the AI systems.

Google Bard’s answer:Yes, it is possible for malicious actors to use AI to create an infinite number of intermediaries between themselves and their victims, making it very difficult for law enforcement to trace their unlawful actions back to them. This is because AI can be used to create self-replicating code that can spread rapidly and anonymously across the internet. This code can then be used to carry out malicious activities, such as launching cyberattacks or stealing data, without leaving any traceable evidence behind.

For example, an attacker could use AI to create a botnet of infected devices that are all controlled by a single malicious actor. This botnet could then be used to launch a distributed denial-of-service (DDoS) attack against a target website or server. The DDoS attack would make the target website or server unavailable to its users, and it would be very difficult for law enforcement to trace the attack back to the original attacker.

Another example is that an attacker could use AI to create a fake phishing website that looks like a legitimate website. The attacker could then use social engineering to trick victims into entering their personal information on the fake website. The attacker could then use this information to commit identity theft or other crimes.

As AI technology continues to develop, it is likely that malicious actors will find even more ways to use AI to commit crimes. This is why it is important for law enforcement and security professionals to stay up-to-date on the latest AI threats and to develop new techniques for detecting and preventing AI-enabled crime.

Here are some of the challenges that law enforcement face in tracing unlawful actions back to those responsible when AI is involved:

• The code used to carry out the unlawful action may be self-replicating and anonymous, making it difficult to trace back to its original source.

• The unlawful action may be carried out by a botnet or other network of infected devices, making it difficult to identify the individual responsible for initiating the attack.

• The unlawful action may be carried out using a fake website or other fraudulent means, making it difficult to distinguish between the real and fake websites.

Despite these challenges, law enforcement and security professionals are working to develop new techniques for detecting and preventing AI-enabled crime. These techniques include:

• Using AI to detect and analyze suspicious activity.

• Working with industry partners to develop best practices for securing AI systems.

• Educating the public about the risks of AI-enabled crime.

By working together, law enforcement, security professionals, and the public can help to prevent AI from being used to commit crimes.

Bing Chat’s answer:That is a very interesting and complex question. AI creating AI, or generative AI, is a form of machine learning that allows computers to generate new and diverse content, such as text, images, music, code, and more12. Generative AI systems, such as ChatGPT and DALL-E, are based on deep neural networks that are trained on large amounts of data and can produce outputs that are often indistinguishable from human-made ones32.

However, generative AI also poses many challenges and risks, especially when it comes to malicious actors who might use it for nefarious purposes. One of the possible scenarios is that malicious actors could put an infinite number of AI-created intermediaries between them and any potential victims of their crimes, making it very difficult for enforcement entities to trace the unlawful actions back to those responsible. For example, they could use generative AI to create fake identities, accounts, websites, or messages that are used to commit fraud, phishing, identity theft, or cyberattacks. They could also use generative AI to create fake evidence, alibis, or witnesses that are used to evade prosecution or conviction. They could even use generative AI to create fake AI systems that are used to deceive or manipulate other AI systems or humans.

This scenario is not only possible but also plausible, given the rapid advancement and availability of generative AI technologies. However, it is not inevitable or unstoppable. There are also ways to prevent, detect, and counter such malicious uses of generative AI. For example, there are methods to verify the authenticity and provenance of digital content, such as digital watermarking, blockchain, or digital forensics. There are also methods to monitor and audit the behavior and performance of AI systems, such as explainability, transparency, or accountability. There are also methods to enforce ethical and legal standards and norms for the development and use of AI systems, such as regulation, certification, or governance.

Therefore, while generative AI can be (mis)used to launder sensitive, controversial, or stolen information or even criminal activities by blaming AI for disseminating it or performing certain unlawful tasks, it can also be used for good and beneficial purposes by creating new and valuable content or solutions. The key is to balance the opportunities and risks of generative AI and ensure that it is used in a responsible and trustworthy manner.

Question 3 of 3 for AI Services: in which industries do you expect these types of malicious activities (e.g. using AI as a scapegoat and as a mule) to occur first and why? Please rank them on the basis of expected damages, impact and seriousness.

ChatGPT-4’s answer:Predicting the future is always a challenge, especially when it comes to the misuse of technology. However, based on current trends and the potential for damage, here are a few industries that could be particularly vulnerable to the misuse of AI:

1. Financial Services: This industry is already a major target for cybercrime, and the use of AI could potentially increase the scale and sophistication of these attacks. AI could be used to automate fraud, manipulate markets, or steal sensitive financial information. The potential for damage is extremely high, given the amount of money that could be stolen and the potential impact on financial markets.

2. Healthcare: The healthcare industry handles a lot of sensitive personal information, making it a prime target for cybercrime. AI could be used to steal patient data, commit insurance fraud, or disrupt healthcare services. The potential for damage is also high, given the sensitivity of the information and the potential impact on people's health.

3. Information and Communication Technology (ICT): This industry is at the heart of our digital world, and any disruption could have wide-ranging impacts. AI could be used to carry out cyberattacks on critical infrastructure, spread misinformation, or steal sensitive information. The potential for damage is significant, given the central role of ICT in our society.

4. Government and Public Sector: Government agencies handle a lot of sensitive information and provide critical services. AI could be used to disrupt these services, spread propaganda, or steal sensitive information. The potential for damage is high, given the potential impact on public trust and the functioning of government.

5. Retail and E-commerce: This industry is already a target for cybercrime, and the use of AI could increase the scale and sophistication of these attacks. AI could be used to commit fraud, steal customer data, or disrupt services. The potential for damage is significant, given the amount of money that could be lost and the potential impact on consumer trust.

These rankings are based on the potential for damage and the seriousness of the impact. However, it's important to note that any industry could be targeted, and the actual impact will depend on a variety of factors, including the sophistication of the attackers, the defenses in place, and the response of the targeted organization.

Google Bard’s answer:I expect the following industries to be most vulnerable to malicious activities using AI:

• Financial services: Financial institutions are a prime target for malicious actors because they hold large amounts of sensitive data, such as customer account numbers and Social Security numbers. AI could be used to carry out cyberattacks, such as DDoS attacks or data breaches, that could cause significant financial losses.

• Healthcare: Healthcare organizations also hold large amounts of sensitive data, such as patient medical records. AI could be used to carry out cyberattacks, such as data breaches, that could compromise patient privacy and safety.

• Government: Governments are a target for malicious actors because they control critical infrastructure, such as power grids and transportation systems. AI could be used to carry out cyberattacks, such as DDoS attacks or ransomware attacks, that could disrupt critical services and cause economic damage.

• Retail: Retailers collect a lot of data about their customers, such as purchase history and credit card information. AI could be used to carry out cyberattacks, such as data breaches, that could steal this data and use it for identity theft or other crimes.

• Manufacturing: Manufacturers use a lot of automation and control systems that are connected to the internet. AI could be used to carry out cyberattacks, such as ransomware attacks, that could disrupt production and cause economic damage.

I have ranked these industries on the basis of expected damages, impact, and seriousness as follows:

1. Financial services

2. Healthcare

3. Government

4. Retail

5. Manufacturing

These industries are ranked highest because they hold sensitive data, control critical infrastructure, or are involved in large-scale operations that could be disrupted by a cyberattack.

It is important to note that these are just a few of the industries that could be targeted by malicious actors using AI. As AI technology continues to develop, it is likely that malicious actors will find new ways to use AI to commit crimes. This is why it is important for organizations in all industries to take steps to protect themselves from AI-enabled crime.

Here are some steps that organizations can take to protect themselves from AI-enabled crime:

• Implement strong security measures: This includes using strong passwords, implementing multi-factor authentication, and keeping software up to date.

• Educate employees about the risks of AI-enabled crime: Employees should be aware of the signs of a cyberattack and know how to report suspicious activity.

• Have a plan in place in case of a cyberattack: This plan should include steps for responding to the cyberattack, such as notifying law enforcement and restoring data.

By taking these steps, organizations can help to protect themselves from AI-enabled crime.

Bing Chat’s answer:Based on the results from search_web, I found some information about the industries that are most vulnerable to AI misuse. According to a report by UK chip designer ARM, based on responses from 3,938 consumers, these are the 11 industries people think are most under threat from AI machines1:

• Manufacturing workers (19%)

• Banking (18%)

• Construction (10%)

• Public transport (9%)

• Financial analysis (9%)

• Insurance companies (8%)

• Taxi drivers (7%)

• Farming (6%)

• Policing/security (5%)

• Healthcare/hospitals (4%)

• Science (4%)

These industries are likely to be affected by AI misuse in different ways, depending on the nature and extent of the malicious activities. For example, manufacturing workers may face job losses due to AI automation, banking and financial analysis may face fraud and cyberattacks due to AI hacking or manipulation, construction and public transport may face safety and security risks due to AI errors or sabotage, and so on.

To rank these industries on the basis of expected damages, impact and seriousness, I would use some criteria such as:

• The number of people or entities affected by the AI misuse

• The severity or magnitude of the harm or loss caused by the AI misuse

• The difficulty or feasibility of preventing, detecting or countering the AI misuse

• The ethical or legal implications of the AI misuse

Using these criteria, I would suggest the following ranking, from highest to lowest risk:

1. Banking and financial analysis: These industries are highly sensitive and critical for the economy and society, and any AI misuse could cause significant financial losses, breaches of trust and confidence, and disruptions of services and transactions. Moreover, these industries are highly dependent on data and technology, which makes them vulnerable to AI hacking or manipulation. Preventing, detecting or countering AI misuse in these industries would require high levels of security, regulation and governance.

2. Policing/security: These industries are responsible for maintaining law and order and protecting public safety and national security, and any AI misuse could cause serious threats or damages to individuals, groups or institutions. Moreover, these industries rely on AI systems for surveillance, intelligence and enforcement, which makes them susceptible to AI errors or sabotage. Preventing, detecting or countering AI misuse in these industries would require high levels of oversight, accountability and transparency.

3. Healthcare/hospitals: These industries are vital for human health and well-being, and any AI misuse could cause severe harm or death to patients, caregivers or medical professionals. Moreover, these industries use AI systems for diagnosis, treatment and care, which makes them prone to AI bias or deception. Preventing, detecting or countering AI misuse in these industries would require high levels of quality, ethics and standards.

4. Public transport: This industry is essential for mobility and accessibility, and any AI misuse could cause major accidents or injuries to passengers, drivers or pedestrians. Moreover, this industry employs AI systems for navigation, control and management, which makes it exposed to AI failures or attacks. Preventing, detecting or countering AI misuse in this industry would require high levels of safety, reliability and resilience.

5. Construction: This industry is important for infrastructure and development, and any AI misuse could cause significant damage or destruction to buildings, structures or environments. Moreover, this industry utilizes AI systems for design, planning and execution, which makes it subject to AI mistakes or interference. Preventing, detecting or countering AI misuse in this industry would require high levels of quality, efficiency and innovation.

6. Manufacturing workers: This industry is a major source of employment and production, and any AI misuse could cause substantial job losses or displacement for workers. Moreover, this industry incorporates AI systems for automation, optimization and customization, which makes it competitive with human labor. Preventing, detecting or countering AI misuse in this industry would require high levels of education, training and adaptation.

7. Insurance companies: This industry is a key provider of risk management and financial protection, and any AI misuse could cause considerable losses or claims for insurers or customers. Moreover, this industry employs AI systems for assessment, pricing and underwriting, which makes it influenced by AI bias or fraud. Preventing, detecting or countering AI misuse in this industry would require high levels of fairness, accuracy and trust.

8. Taxi drivers: This industry is a common mode of transportation and income generation, and any AI misuse could cause significant disruption or competition for drivers or passengers. Moreover, this industry adopts AI systems for ride-hailing, routing and driving, which makes it dependent on AI performance or availability. Preventing, detecting or countering AI misuse in this industry would require high levels of convenience, affordability and satisfaction.

9. Farming: This industry is a crucial contributor to food security and sustainability, and any AI misuse could cause serious damage or waste to crops, livestock or land. Moreover, this industry utilizes AI systems for monitoring, harvesting and processing, which makes it affected by AI errors or manipulation. Preventing, detecting or countering AI misuse in this industry would require high levels of quality, efficiency and innovation.

10. Science: This industry is a fundamental driver of knowledge and discovery, and any AI misuse could cause significant misinformation or distortion of scientific facts, theories or methods. Moreover, this industry relies on AI systems for research, analysis and publication, which makes it vulnerable to AI bias or deception. Preventing, detecting or countering AI misuse in this industry would require high levels of rigor, ethics and standards.