Generative AI (Artificial Intelligence): Benefits and Challenges – Explained, pointwise

Introduction

Since its release, ChatGPT has received a lot of attention. While, the users are marvelling at its ‘human-like’ responses, technology experts are debating the potential applications and concerns associated with Generative AI (Artificial IntellIgence). Generative AI has the potential to revolutionise almost every field of human activity. However, the possibility of misuse of the technology and loss of skilled and semi-skilled jobs has prompted calls for more cautious approach in the development of the Generative AI.

What is Generative AI?

Generative AI uses Artificial Intelligence and Machine Learning algorithms to enable machines to generate new content (machine generated). Systems use previously created content, such as text, audio, video, images, and code. The term ‘Generative’ refers to the ability of the models to learn how to create new data rather than simply recognising it. For example, a generative model may learn how to generate images that resemble faces given a set of parameters (such as the eyes, hair, or skin colour etc.). The content (text, image etc.) generated by AI is so ‘authentic’, that it is difficult to distinguish whether the content has been generated by human or computer.

Source: WEF. The above image has been created by Generative Artificial Intelligence developed by Midjourney Labs. The text prompt to generate the image was ‘A technical illustration of a woman sitting behind a desktop computer on a long table, isometric view, 3D rendering, realistic…’

What are the applications of Generative AI?

The applications of Generative AI are wide and still evolving.

Motion Picture Industry: Applications of Generative Artificial Intelligence in the movie industry is wide. It can utilized to alter the background/landscape according to the need (rather than wait for required conditions to exist e.g., a movie scene requiring cloudy weather can be shot under any weather conditions, and the background can be altered later on using AI). Images or videos of Actors at various ages are also possible with Generative AI technology. By using face synthesis and voice cloning, artist’s/actor’s original voice can be matched with a lip-sync. This will also help in archiving artefacts after restoration for future references.

Search Engine Services: Generative Artificial Intelligence has the capability to take search engine services to the next level, e.g., Text to Image translation may be utilized to provide search results. It can also be used to produce realistic photographs from textual descriptions of objects like birds and flowers.

Source: WEF. Images created by Midjourney through Generative Artificial Intelligence using Text Prompt.

Security Services: Generative Artificial Intelligence can create front-on photos from photos taken at different angles and vice versa for face verification or face identification systems. Such systems can be deployed at airports, international border check-points etc.

Healthcare: Semantic-Image-to-Photo Translation can convert inputs that are semantic images or sketches to photo-realistic images e.g., if X-ray or any CT scan images can be converted to real images, diagnosis can be much more accurate.

Advertising: Generative AI can create new advertisements based on existing ones, making it easier for companies to reach new audiences.

Location Services: This involves converting satellite images to map views. This can be a huge step towards venturing into unexplored geographic locations.

The possible applications of Generative AI are still being explored and can expand considerably as the technology evolves further. It can expand to fields like education, content creation, banking among others.

What are the benefits of Generative AI?

Increased Efficiency: Generative Artificial Intelligence can be used to automate tasks that would otherwise require manual labor. This can help businesses save time and money, as well as increase efficiency e.g., it can be used to generate images and videos quickly and accurately, which can be used in marketing campaigns or other projects.

Improved Quality: Generative Artificial Intelligence can help improve the quality of content generated. It can be used to create high-quality images and videos that are more visually appealing than those created manually. Additionally, it can be used to generate text that is more accurate and relevant than text created by humans.

Faster Results: Generative Artificial Intelligence can help businesses get results faster than they would with manual labor. It can create images and videos in a fraction of the time it would take a human to do the same task. This can help businesses get their projects done at a much faster rate.

Cost Savings: By automating tasks, businesses can reduce their labor costs and save money. Additionally, it can help businesses reduce costs associated with creating content, such as images and videos.

Improved Decision Making: By using Generative AI, businesses can generate data that can be used to make better decisions e.g., it can be used to generate data that can be used to make decisions about marketing campaigns or product development. Applications in the medical field can help in better diagnosis.

Increased Creativity: Businesses can generate new ideas and concepts that can be used to create new products or services.

Improved Customer Experience: Businesses can generate content that is more accurate and relevant to their customers. This can help businesses create a better customer experience and increase customer satisfaction.

What are the concerns associated with Generative AI?

Accuracy: Despite the advancements, the Generative AI technology is not fool proof and produce erroneous content. The Machine Learning Algorithms depend on the quality of the input data. Erroneous or inaccurate data can generate inaccurate results.

Increase Biases: Generative Artificial Intelligence systems can perpetuate and amplify existing biases. If the models are trained on biased, non-inclusive data, they will generate biased outputs, such as offensive or discriminatory language, demeaning and degrading imagery, and prejudicial content. A rights-group in the US pointed out the example of an AI-based generative imagery programme showing images of only white men for the prompt ‘CEO’.

Malicious Purposes: Generative AI systems can create content for malicious purposes, such as deepfakes, disinformation, and propaganda. It can also generate offensive or inappropriate content. Nefarious actors may use AI-generated media to manipulate people and influence public opinion. It can be misused by enemy States, or non-State actors to destabilise domestic peace by spreading misinformation.

Low Quality Output: It may also produce low-quality and less accurate information, specifically in the context of complex engineering and medical diagnosis.

Concern over Data Privacy: Data privacy issues can arise from using generative AI in different industries, such as healthcare, since it involves collecting private information about individuals.

Limitations in Creativity: AI uses past data as a template for future work. It means that the output produced by Generative AI is usually based on something that has already happened rather than anything genuinely creative. In short, AI systems lack creativity, originality and human ingenuity. AI Systems cannot generate new ideas by themselves, they can only make associations based on the data fed into them by humans.

Issues Related to Copyright: It can be challenging to determine who is responsible for the content generated by a Generative AI system. The acquisition and consent model around the training data and intellectual property issues make it difficult to hold anyone accountable for any harm resulting from its use. In addition, there are concerns related to use of copyrighted content to train AI systems. The work derived from such content can have copyright implications. Getty Images has sued Stable Diffusion in the London High Court, accusing them of using its images illegally.

Risk of Unemployment: Although it is too early to make certain judgements, there is a risk that generative AI could contribute to unemployment in certain fields. This could happen if generative AI automates tasks or processes previously performed by humans, leading to the displacement of human workers.

Environmental Concerns: AI systems require a lot of computing power. This has implications for environments, in terms of energy consumed in operating AI systems. An analyst pointed out that training a transformer model just once with 213 million parameters can emit carbon emissions equivalent to 125 air-flights between New York and Beijing. GPT3 has 175 billion parameters, so its emissions would have been much larger.

What should be done going ahead?

First, To address bias and fairness, researchers can use techniques such as de-biasing and fair representation learning, which can help to remove biases present in the training data.

Second, Researchers can also use techniques such as counterfactual data generation, which can help to generate more diverse and representative training

Third, There is need to add rigour and responsibility to developing AI technology, develop and enforce ethical guidelines, conduct regular audits for fairness, identify and address biases, and protect privacy and security.

Fourth, There is need to add adequate policy, regulation, awareness, and education guardrails to develop and use Generative AI services ethically and responsibly. China has proposed a policy for the same. Some measures include requirement for the users of Generative AI to ensure that any doctored content using the technology is explicitly labelled and can be traced back to its source. The regulation also mandates people using the technology to edit someone’s image or voice, to notify and take the consent of the person in question.

Fifth, Intellectual property law must find a way to protect artists from copies that erode the value of their original work, but at the same time encourage them to continue to be inspired by others. The US Copyright Office has already declared that AI generated art is not entitled to intellectual property protection as it lacks the ‘nexus between the human mind and creative expression’, which is necessary to invoke copyright protection.

Conclusion

The Generative AI is a revolutionary technological development. However, as is the case with every new technology, it has several associated concerns. A pragmatic approach is necessary that can minimize the negative impacts of technology. A cooperation at a global level will be required to establish the norms and standards, as well as checking misuse of the technology that can transcend national boundaries.

Syllabus: GS III, Awareness in the fields of IT and Computers

Source: Mint, Mint, The Hindu, WEF