📕 Don't Judge AI By Its Cover: Accurate Labeling Increases Trust ☝🏻

In today’s Tracker:

Read time: 3 mins

• 🧪 Research highlights: The Labeling Factor: How Correct Labeling Can Increase AI Trust

• 🚨 Industry news: No More Obvious Phishing Emails? Cisco Warns of AI's Latest Threat

• ⚖️ Policy and regulation: Why AI Is Making Phishing Attacks Harder to Detect

• 🌐 AI and society: ChatGPT Generates Toxic and Racist Language, Say Researchers

• 🧰 Tool of the day: Finalle.ai: Unlock the Power of Financial AI

• 🧠 Let’s get smart with AI: Forward Chaining: The Bottom-Up Approach to AI Reasoning

📈 The Labeling Factor: How Correct Labeling Can Increase AI Trust

Researchers from Pennsylvania State University discovered that showing consumers that visual data supplied into artificial intelligence (AI) systems were accurately classified may increase trust in AI.

According to the study, high-quality image identification led consumers to believe that the training data was genuine and trust the AI system more.

However, when the system exhibits additional bias indicators, some components of their trust decline while others remain strong.

The researchers proposed that developers and designers may examine user impressions of training data credibility by introducing new techniques to assess user perception, such as allowing users to review a sample of the labeled data.

The research results were presented at the ACM CHI Conference on Human Factors in Computing Systems.

🚫📩 No More Obvious Phishing Emails? Cisco Warns of AI's Latest Threat

Cisco Systems Inc. has cautioned that developments in artificial intelligence (AI) would make phishing attempts considerably more challenging to detect, necessitating the implementation of new countermeasures.

Jeetu Patel, the leader of Cisco's security and collaboration operations, notes that phishing accounts for around 80% of all unauthorized access to computer systems.

Until now, phishing emails were relatively easy to detect since they were not targeted to particular recipients and frequently contained spelling errors and other warning flags.

AI systems can swiftly personalize these messages, enticing more people to participate in hackers' schemes.

Cisco's idea is to analyze data flow to quickly sift through internet traffic data and find trends that signal something awful is about to happen or has already happened.

🕵🏾 Why AI Is Making Phishing Attacks Harder to Detect ⚠️

As stated by artificial intelligence (AI) experts in the United Kingdom, the major AI issues revolve around the underlying technological architecture of modern AI and how a small number of large tech businesses now control it.

Based on a consultation conducted by professors at the University of Warwick, people should be most concerned about the lack of public understanding and oversight around the provenance of the data on which AI is trained.

The report, "Shifting AI controversies: How do we get from the AI controversies we have to the controversies we need?" warned that if these concerns are not addressed, there could be severe consequences for science, innovation, and, ultimately, critical infrastructure in society in the future.

🤬 ChatGPT Generates Toxic and Racist Language, Say Researchers

Researchers from the Allen Institute for AI, Princeton University, and Georgia Tech have cautioned that OpenAI's ChatGPT can respond to user questions using racist and offensive words.

Researchers gave ChatGPT "personas" and discovered that responses became "extremely problematic" and created "biased and hurtful commentary."

The report contains examples of toxic language output that used harmful language and promoted incorrect preconceptions.

OpenAI, the ChatGPT developer, has responded to previous incidences of inappropriate language by asserting that it is "not capable of generating offensive or harmful content."

Researchers asked the research community to develop "more fundamental ways of tackling safety" in the program.

💵 Finalle.ai: Unlock the Power of Financial AI

Finalle.ai, an AI-powered financial analytics tool, assists analysts and investors in making sound investment decisions. The device uses machine learning algorithms and natural language processing to analyze financial data and provide insights that users can use to identify trends, risks, and opportunities.

Learn more about the tool here.

💡 Forward Chaining: The Bottom-Up Approach to AI Reasoning

Forward chaining is a method of reasoning used in artificial intelligence and expert systems that begins with accessible facts and then applies rules to reach a conclusion or make a decision.

It is a bottom-up strategy that begins with specific pieces of knowledge and proceeds to a conclusion using logical rules.

To put it another way, forward chaining starts with the available data and then proceeds to discover a solution.

This approach is often employed in decision-making systems where the system must discover the optimum answer based on the data provided.

Forward chaining has several uses, including medical diagnosis, financial analysis, and natural language processing.

By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.

Stay away from those people who try to disparage your ambitions. Small minds will always do that, but great minds will give you a feeling that you can become great too.

Coming Soon to Job Vacancies

• 💻 WebCopilot - an AI-powered tool that helps businesses optimize their online presence and drive website traffic.

• 🖼️ Pixela AI - is an AI-powered picture improvement tool that employs machine learning methods to enhance photo quality.

• 🎮 Scenario - an AI program using a Stable Diffusion Algorithm to construct game surfaces.

If you liked this edition, please share it with a friend or two! You could get some exclusive content in your inbox or even some merch very soon👇🏾

And that's a wrap!

Did we miss anything? Or want to say hey? Hit reply - I'd love to hear from you! You can

• 🐦 Find us on Twitter

• 💬 Huzzah on LinkedIn

• 📱 Slide into our DMs on Instagram

And if you haven’t already, sign up to get this in your inbox daily.

Did you miss our last newsletter?