Learn the Most Important AI Models
Grok 4: Problematic AI in Swedish School Environment
Reading time: approx. 10 min
After exploring the more established AI models from OpenAI, Anthropic and Google, it is time to look at a newer and more controversial player on the AI scene: Grok 4. Developed by xAI, Elon Musk's AI company, Grok 4 was launched on July 9, 2025. It stands out with its ability to access real-time information via the platform X (formerly Twitter), a distinct "rebellious" personality and the new image AI engine Aurora. However, the characteristics and recent controversies make Grok 4 potentially problematic to use in Swedish school environments without very careful consideration.
What you will learn
- What distinguishes Grok 4 and the connection to platform X.
- Grok's new image AI engine Aurora and its functions.
- Grok's distinct personality and impact on interactions.
- Important considerations regarding data protection, security and ethical use in school environments, due to controversies and risks of disinformation.
- Grok's language support for Swedish.
The Basics: What is Grok 4 and the Ecosystem?
Grok 4 is the latest flagship model from xAI, Elon Musk's AI company. xAI was founded to develop AI that is "maximally truth-seeking and questioning" and released Grok-1 as open source in March 2024.
One of Grok 4's most prominent features is direct access to information from platform X and external web search in real-time. This gives it a unique advantage when it comes to answering questions about current events and trends, something other models can struggle with as they usually have a "cut-off date" for their training data.
Licensing and Availability
To access Grok 4 and the real-time features currently requires a paid subscription on platform X. This can be X Premium+ (approximately 300/month), which offers higher usage limits and performance. Bulk discounts may be available for larger organizations.
"Rebellious" Personality and Humor
xAI has deliberately designed Grok to have a specific "personality". It is described as having a "rebellious streak" and likes to answer questions with a touch of humor and sarcasm. It is meant to be less "politically correct" and more inclined to give more direct answers, which can be both an advantage and a disadvantage depending on use case and context, especially in school environments.
Strengths: What is Grok 4 Good At?
Real-time knowledge and current events: Grok 4's direct access to information from X and external web search makes it exceptional at answering questions about the latest news, trends and events developing in real-time. This is a unique strength for getting quick updates on current topics.
- Practical example: "What are the latest headlines about the climate conference happening right now?" or "Summarize the most discussed topics on X today regarding the Swedish election."
Image generation with Aurora: Grok 4 is powered by the new image AI engine Aurora, which is an autoregressive MoE model (Mixture-of-Experts). Aurora is being rolled out gradually and enables image generation directly from text descriptions. Since March 2025, users can also upload and edit images directly in Grok.
- Practical example: "Generate an image illustrating how AI can be used in a history lesson" or "Edit this image by adding a sunset in the background."
Unique personality and tone: Grok can deliver answers with a distinct humorous or sarcastic tone, which can be useful for creative writing or role-playing in certain specific contexts.
Swedish Language Handling
Grok supports over 25 languages, including Swedish. It is trained on a large amount of text data and can generally understand and respond in Swedish. However, tests show that the tone in Swedish responses can sometimes be irregular or not entirely consistent with Grok's intended personality in English. Teachers should be aware of this when using it.
Important Considerations for School Environments (Data Protection, Security & Ethics)
Grok's unique characteristics and rapid development have led to several discussions and controversies that teachers should be especially aware of, particularly in a Swedish school environment:
Reliability, disinformation and source criticism: Since Grok takes information directly from X, which can contain unreliable sources, "fake news", polarized opinions and disinformation, source criticism is extremely important. Platform X has been subject to EU investigations and blockades related to inadequate moderation. Additionally, the European Parliament and EU Commission actively warn about the recent wave of disinformation, including the use of generative AI to spread "deepfakes" and unreliable narratives (for example related to the Ukraine war). Grok's real-time capability, combined with sometimes inadequate moderation on X, increases the risk that the model references and spreads unreliable information. Real-time search may also filter out certain topics after the latest scandal.
- Solution: Use Grok to get an overview of a discussion or the latest trends, but always verify facts via reliable, external sources before information is used in teaching. Teach students to be critical of AI-generated information, especially that coming from real-time feeds and platforms with inadequate moderation.
Data protection and privacy: The EU's e-safety authority (eSafety Commissioner) has expressed warnings about limited moderation and the potential for unwanted content via public X threads that AI can reference. As a commercial AI service, the same data protection questions apply as for other platforms.
- Solution: Avoid entering personal data or sensitive student data in Grok. Your school's policy for AI use must be followed.
Personality, ethics and controversies: After controversies in July 2025 regarding antisemitic content and inadequate moderation, xAI has promised stricter system prompts. Despite this, the problems with Grok's "rebellious" tone and the risk of inappropriate content are not completely resolved. This has led to extensive discussions in the media.
- Solution: Be aware of Grok's tone and whether it is suitable for the intended use. Red-teaming (security testing) has shown that Grok 4 requires extra "security prompting" (clear security instructions) for use in sensitive environments like schools. It can be a good model to use to discuss AI "personality", how AI can be biased, and the importance of responsible design and moderation.
Implementation in the Classroom
- Current events and discussions (with source criticism): Use Grok to get quick insights into what is being discussed in real-time around current topics, which can be a starting point for classroom discussions. Remember to always critically evaluate the information.
- Exercise in source criticism and media literacy: Grok can be an excellent tool for demonstrating the importance of source criticism and media literacy, as it can deliver quick answers that are potentially unreliable, biased or controversial. Students can analyze Grok's answers to identify potential disinformation or biased presentations.
- Creative writing and AI personality: For older students who understand AI limitations and can handle sarcasm, Grok can be used to explore different text styles or to discuss AI "personality" and potential effects on content.
Next Steps
In the next module we will take a closer look at DeepSeek, an interesting player among the open source models, known for its performance in coding and reasoning.