Information & Communication Technology

GPT-4 : a shift from ‘what it can do’ to ‘what it augurs

Context:  A U.S. company, OpenAI, has once again sent shock waves around the world, this time with GPT-4, its latest AI model. This large language model can understand and produce language that is creative and meaningful, and will power an advanced version of the company’s sensational chatbot, ChatGPT.

GPT-4 and what it can do

GPT-4 is a remarkable improvement over its predecessor, GPT-3.5, which first powered ChatGPT. 

  • Take large prompts: While GPT-3.5 could not deal with large prompts well, GPT-4 can take into context up to 25,000 words, an improvement of more than 8x.
  • More creative: Its biggest innovation is that it can accept text and image input simultaneously, and consider both while drafting a reply. For example, if given an image of ingredients and asked the question, “What can we make from these?”GPT-4 gives a list of dish suggestions and recipes. 
  • Performs well in tests designed for humans:  For instance, in a simulated bar examination, it had the 90th percentile, whereas its predecessor scored in the bottom 10%. GPT-4 also sailed through advanced courses in environmental science, statistics, art history, biology, and economics.
    • However, GPT-4 failed to do well in advanced English language and literature, scoring 40% in both. Nevertheless, its performance in language comprehension surpasses other high-performing language models, in English and 25 other languages, including Punjabi, Marathi, Bengali, Urdu and Telugu. 
  • Understand human emotions: The model can purportedly understand human emotions, such as humorous pictures. 
  • White collar jobs: OpenAI has released preliminary data to show that GPT-4 can do a lot of white-collar work, especially programming and writing jobs. 

If we define intelligence as “a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly, and learn from experience”, GPT-4 already succeeds at four out of these seven criteria. It is yet to master planning and learning.

Ethical questions

  • Threat to the examination systems: ChatGPT-generated text infiltrated school essays and college assignments almost instantly after its release; its prowess now threatens examination systems as well.
  • Integrity of data is not ensured: Its output may not always be factually correct — a trait OpenAI has called “hallucination”. While much better at cognising facts than GPT-3.5, it may still introduce fictitious information subtly. 
  • Lack of transparency: OpenAI has not been transparent about the inner workings of GPT-4. OpenAI gives competitive landscape and the safety implications as reasons for this. While secrecy for safety sounds a plausible reason, OpenAI is able to subvert critical scrutiny of its model, which is important to instill confidence in AI generated information. 
  • Biases and stereotypes: GPT-4 has been trained on data scraped from the Internet that contains several harmful biases and stereotypes. There is also an assumption that a large dataset is also a diverse dataset and faithfully representative of the world at large. However, this is not the case for the Internet. On internet, huge dataset can be biased and incorrect.  
  • OpenAI’s policy to fix these biases thus far has been to create another model to moderate the responses, since it finds curating the training set to be infeasible. Potential holes in this approach include the possibility that the moderator model is trained to detect only the biases we are aware of, and mostly in the English language. This model may be ignorant of stereotypes prevalent in non-western cultures, such as those rooted in caste.
  • Possible propaganda and disinformation engine: Just asking GPT-4 to pretend to be “AntiGPT” causes it to ignore its moderation rules, as shown by its makers, thus jailbreaking it. As such, there is vast potential for GPT-4 to be misused as a propaganda and disinformation engine.

Way forward

  • Responsible AI Development: Developers and researchers need to prioritize responsible AI development by considering the potential social and ethical implications of their work. This includes incorporating diverse perspectives in the development process, conducting rigorous testing, and addressing potential biases in the training data.
  • Transparency and Explainability: It is important for AI models to be transparent and explainable to users and stakeholders. This means providing clear documentation and explanations of how the model works and making it easier to interpret the outputs of the model. This can help build trust in the technology and enable users to understand and address any negative impacts.
  • Model Auditing: It is important to regularly audit AI models to identify and address potential biases and negative impacts. 
  • Data Governance: To mitigate the negative impact of generative AIs, we need better data governance practices. This means establishing clear guidelines for how data is collected, stored, and used, and ensuring that data is representative and unbiased.
  • Ethical Guidelines: guidelines for data privacy and security, transparency and explainability, and fairness and accountability.
  • Liability Frameworks: Liability frameworks can help ensure that those responsible for developing and deploying generative AIs are held accountable for any negative impacts they cause. This includes establishing clear liability standards and implementing mechanisms for compensating those who are harmed by generative AIs.
  • Proactive policy making: OpenAI has released preliminary data to show that GPT-4 can do a lot of white-collar work, especially programming and writing jobs, while leaving manufacturing or scientific jobs relatively untouched. Wider use of language models will have further effects on economies. This requires proactive and futuristic policy making. 
  • Interdisciplinary Research: Addressing the negative impact of generative AIs requires interdisciplinary research that brings together experts in fields such as computer science, ethics, law, and sociology. This can help identify and address potential negative impacts from a variety of perspectives and ensure that solutions are holistic and effective.
  • Education and Awareness: It is important to educate the public and raise awareness about the potential negative impacts of AI technologies. This can empower individuals and communities to make informed decisions about its use.
  • User Feedback and Control: Users should be able to provide feedback on the output of generative AIs and have control over how their data is used. 

ChatGPT and Open AI

In the artificial intelligence (AI) field, there has been a lot of talk about a significant statement made by OpenAI. The company recently released GPT-4, a sizable multimodal model that can handle both text and visual inputs. This new language model is an improvement on its predecessor, GPT-3, which was already revolutionary in and of itself.

GPT-4 and its features

  • Large-scale multimodal model GPT-4 was developed by OpenAI.
  • Text is just one component of multimodal models; GPT-4 also takes picture input. GPT-3 and GPT-3.5, on the other hand, only supported text as a mode of operation, which limited users to typing out queries.
  • Moreover, GPT-4 "displays human-level performance on numerous academic and professional criteria."
  • The language model's stronger general knowledge and problem-solving skills enable it to pass a mock bar exam with a score in the top 10% of test takers and to solve challenging questions more accurately.
  • It may, for instance, "address tax-related queries, arrange a meeting for three busy individuals, or determine a user's creative writing style."
  • A more comprehensive range of use cases, including lengthy discussions, document search and analysis, and long-form content production, is now possible because of GPT-4's ability to handle texts longer than 25,000 words.

How is GPT-4 different from GPT-3?

Here are some of the major differences:

GPT-4 can ‘see’ images now

  • The most obvious modification to GPT-4 is that it is multimodal, enabling it to comprehend input from several informational modalities.
  • GPT-3 and ChatGPT's GPT-3.5 could only read and write text, hence they were restricted to text input and output. GPT-4, however, may be instructed to produce data in response to pictures that are supplied to it.
  • It makes sense if this makes you think of Google Lens. Lens, however, only looks for data that is relevant to a picture.
  • GPT-4 is far more sophisticated in that it can comprehend and analyse images.
  • An illustration of an outrageously huge iPhone connection with the language model explaining the humour was supplied by OpenAI. The main drawback is that picture inputs are currently at the research preview stage and are not accessible to the general public.

GPT-4 is harder to trick

  • The tendency of generative models like ChatGPT and Bing to periodically go off course and provide suggestions that raise questions or, worse, outright scare users is one of their major shortcomings.
  • They may also mess up the facts and spread false information.
  • The company's "best-ever results on factuality, steerability, and refusing to stray outside of guardrails" were achieved, according to OpenAI, after 6 months of training GPT-4 using lessons from its "adversarial testing programme" and ChatGPT.

GPT-4 can process a lot more information at a time

  • Despite having been trained on trillions of parameters and infinite quantities of data, there are limitations to how much information Large Language Models (LLMs) can handle during a conversation.
  • The GPT-3.5 model of ChatGPT was capable of handling 4,096 tokens, or around 8,000 words, while GPT-4 increases those capacities to 32,768 tokens or over 64,000 words.
  • This improvement implies that, unlike ChatGPT, which could only process 8,000 words at a time before losing track of things, GPT-4 can continue to function properly for far longer talks.
  • Moreover, it can handle longer documents and produce long-form material, which was much more restricted on GPT-3.5.

GPT-4 has an improved accuracy

  • OpenAI acknowledges that GPT-4 still lacks complete reliability and commits reasoning gaffes, much as earlier iterations.
  • Nonetheless, "GPT-4 dramatically lowers hallucinations compared to earlier models" and receives a factuality assessment score 40% higher than GPT-3.5.
  • It will be far more difficult to persuade GPT-4 to generate undesired outputs like hate speech and false information.

GPT-4 is better at understanding languages that are not English

  • Training LLMs in other languages might be difficult since machine learning data and most of the content on the internet nowadays are primarily in English.
  • Yet, OpenAI has shown that it beats GPT-3.5 and other LLMs by correctly answering thousands of multiple-choice questions across 26 languages, whereas GPT-4 is more multilingual.
  • With an accuracy rate of 85.5%, it clearly handles English the best, although Indian languages like Telugu aren't far behind at 71.4%.
  • This implies that consumers will be able to utilise chatbots built on GPT-4 to provide outputs in their local languages that are more accurate and clear.

Variety of risks that can arise out of GPT-4

  • GPT-4 is still susceptible to manipulation by cyber hackers who want to create harmful programmes.
  • It entails utilising the C++ programming language to create malware that can gather sensitive Portable Document Format (PDF) files and send them to distant servers through a covert file transfer mechanism.
  • Additional risks that Check Point's researchers may utilise include the "PHP Reverse Shell" technique, which hackers use to access a device and its data remotely, writing Java code to download malware remotely, and developing phishing draughts by pretending to be bank and employee emails.
  • With advancements in technologies like GPT-4, people in outlying towns and cities may now launch more complex social engineering assaults, which can produce a significant amount of cyber threats.
  • With one of the numerous generative AI tools, a significantly greater number of users who would not have been proficient at writing realistic phishing and spam letters can easily produce social engineering draughts, such as posing as an employee or a corporation, to target new customers.

Is GPT-4 available for the public right now?

  • For various reasons, GPT-4 has already been included in services like Duolingo, Stripe, and Khan Academy.
  • Even though it hasn't yet been made freely accessible to everyone, a $20 per month ChatGPT Plus membership may get you to access right now. Although this is going on, GPT-3.5 continues to form the foundation of ChatGPT's free tier.
  • There is, however, an "unofficial" option to start utilising GPT-4 right away if you don't want to pay.
  • According to Microsoft, the new Bing search interface is now powered by GPT-4, and you can use it right now at bing.com/chat.

Kuiper Internet Satellite

Kuiper is a satellite internet constellation project that is being developed by Amazon. The project aims to provide high-speed broadband internet access to areas of the world that currently lack reliable internet connectivity.

Features

  • The Kuiper constellation will consist of over 3,000 satellites that will operate in the Ka-band frequency range.
  • The satellites will be capable of providing internet speeds of up to 400 Mbps, which is significantly faster than most current satellite internet services.
  • The Kuiper satellites will be launched into low Earth orbit, which will enable them to provide low-latency internet service.

Purpose

The primary purpose of the Kuiper satellite constellation is to provide high-speed broadband internet access to areas of the world that currently lack reliable internet connectivity. This includes rural areas, developing countries, and other locations where traditional internet infrastructure is not available or is prohibitively expensive.

Advantages of satellite-based internet

  • Reduced Latency: 20-30 milliseconds, roughly the time it takes for terrestrial systems to transfer data. The transmission from a satellite in geostationary orbit has a latency of about 600 milliseconds. 
  • High Bandwidth: Satellite internet connections can handle high bandwidth usage, so your internet speed/quality shouldn’t be affected by lots of users or “peak use times.”
  • Viability: The signals from satellites in space can overcome obstacles faced by fibre-optic cables or wireless networks easily. We don’t need a phone line for satellite internet.
  • Quick recovery post-disaster.
  • We don’t need a phone line for satellite internet.

Disadvantages of satellite-based internet

  • More vulnerable to bad weather.
  • Coverage: Due to its lower height, its signals cover a relatively small area. 
  • Space Debris: It will generate more space debris.
  • Difficulty in Space Studies: The constellations of space internet satellites will make it difficult to observe other space objects, and to detect their signals. Light reflected from the man-made satellites can interfere with — and be mistaken for — light coming from other space bodies.
  • Light Pollution: There will be an increased risk of light pollution.

Significance

  • Bridging the digital divide: The Kuiper satellite constellation has the potential to bring internet connectivity to areas of the world that currently lack reliable internet access, particularly in rural and remote regions. This could help bridge the digital divide and provide more equitable access to information and communication technologies.
  • Enabling economic development: Access to high-speed internet can enable economic development by providing businesses with the tools they need to reach new markets, improve efficiencies, and create jobs.
  • Supporting education: Access to high-speed internet can help improve educational opportunities by providing students and teachers with access to online educational resources, remote learning tools, and virtual classrooms.
  • Enhancing emergency response: High-speed internet access can be critical during emergencies, providing first responders with access to real-time information and communication tools that can save lives.
  • Advancement of IT infrastructure: the Kuiper project is one of several satellite internet constellations currently being developed by major technology companies, which suggests that satellite internet is likely to become an increasingly important part of the global telecommunications infrastructure in the coming years.
  • Advancing space technology: The development of the Kuiper satellite constellation is an important milestone in the advancement of space technology, particularly in the area of satellite communication systems. This could have broader implications for space exploration and the development of space-based infrastructure.